For Peer Review - Wireless Communications Lab

For Peer Review

1

Evolution and Future Trends of Research in

Cognitive Radio: A Contemporary Survey

Ekram Hossain, Dusit Niyato, and Dong In Kim

Abstract

The cognitive radio (CR) paradigm for designing next generation wireless communications systems is becoming

increasingly popular and different aspects of it are being implemented in currently available wireless systems.

In the last decade, significant amount of research efforts have been made to solve CR challenges and several

standards related to cognitive radio and dynamic spectrum access have been developed. Also, there have been

advances in software-defined radio (SDR) platforms to implement the cognitive radio systems. In this article,

we provide a comprehensive survey on the evolution of CR research covering aspects such as spectrum sensing,

measurements and statistical modeling of spectrum usage, physical layer (PHY) aspects such as waveform and

modulation design, multiple access, resource allocation and power control, cognitive learning, adaptation and self-

configuration, multi-hop transmission and routing, and robustness and security in cognitive radio networks. Also,

state-of-the-art research on the economics of cognitive radio networks, cognitive radio simulation tools, testbeds

and hardware prototypes, cognitive radio applications, and cognitive radio standardization efforts are summarized.

Emerging trends on cognitive radio research and open research challenges related to the cost-effective and large-

scale deployment of cognitive radio systems are outlined.

I. INTRODUCTION AND HISTORICAL NOTES ON COGNITIVE RADIO

For wireless communications, the right to access the spectrum or license is generally defined by

frequency, space, transmission power, spectrum owner (i.e., licensee), type of use, and the duration of

license. Normally, a license is assigned to one licensee, and the use of spectrum by this licensee must

conform to the specification in the license (e.g., maximum transmission power and type of use). In the

traditional spectrum licensing scheme, the license cannot change the type of use or transfer the right to

other licensees. Also, the radio spectrum is licensed for a large region and generally in large chunks.

All these factors in the current command-and-control model for spectrum allocation and assignment limit

the use and result in low utilization of the frequency spectrum. Since the existing and the new wireless

Page 3 of 66

http://mc.manuscriptcentral.com/wcm

Wireless Communications and Mobile Computing

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

2

applications and services are demanding for more transmission capacity and data transmission performance

and hence more radio spectrum, the utilization of the radio spectrum needs to be improved.

To improve the efficiency and utilization of the radio spectrum, the limitations above should be remedied

by modifying the spectrum licensing scheme and adopting a dynamic spectrum management model. The

idea is to make spectrum access more flexible by allowing the unlicensed users to access the radio

spectrum under certain restrictions. Since the legacy wireless systems were designed to operate on a

dedicated frequency band, they are not able to utilize the improved flexibility provided by this spectrum

licensing scheme. Therefore, the concept of cognitive radio emerged the main goal of which is to provide

adaptability to wireless transmission through dynamic spectrum access so that the utilization of the

frequency spectrum can be enhanced without losing the benefits associated with static spectrum allocation.

The cognitive radio is a “smarter radio” in the sense that it can sense channels that contain signals from

a large class of heterogeneous devices, networks, and services. Based on this sensing, the radio will

implement sophisticated algorithms to share the limited-bandwidth channel with other users in order to

achieve efficient wireless communication. In this way, the cognitive radio concept generalizes the idea of

multiple access involving devices in a single homogeneous system to multiple access among devices in

different radio spectrum using different radio transmission techniques and hence different systems (i.e.,

inter-system multiple access as opposed to the more traditional intra-system multiple access) which have

different priorities in accessing the spectrum [1].

Again, the concept of “cognitive radio” can be generalized as a topic within the field of adaptive or

cognitive systems. The cognitive radio refers to a sophisticated radio device that mimics the human brain

in a dynamic spectrum access (DSA) environment to perceive and learn the radio environment to control

and adapt the transmission actions [2]. Such a radio device or terminal has to operate in a networked

environment where many of such devices sense the radio channels in their own geographical environment

and the sensed information is processed (either centrally or distributively) to make a decision on channel

access using appropriate modulation scheme, transmission power and error control algorithms. The wireless

network infrastructure will also have a significant role here, for example, to manage information about

the radio devices and their operating environments, determine the consequences of interference, facilitate

cooperation among the radio devices, etc. Significant challenges exist in the design, analysis, optimization,

development, and deployment of DSA-based cognitive radio networks (CRNs), and these challenges in-

volve both technology and policy (or regulatory) considerations. The major technical challenges include the

following: spectrum sensing, measurements and statistical modeling of radio spectrum usage; waveform,

Page 4 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

3

modulation, and RF front-end design at the physical layer; methods for multiple access, resource allocation

and transmission power control for cognitive radios; distributed learning, adaptation, self-configuration,

cooperation, and coexistence methods for cognitive radios; routing in multi-hop CRNs and cross-layer

optimization; methods for robustness and security in CRNs; wide-area end-to-end performance modeling

and analysis for global internetworking; modeling and understanding the emergent system behavior;

development of network control and management protocols; software abstractions for the radio modules at

the application programming interface (API) level; and development of testbeds and real-world deployment

of CRNs. The challenges in the spectrum policy domain include the following: development of practical

DSA policies that lead to efficient spectrum use, protect the rights of license holders, and maintain the

service quality; development of certification and enforcement policies for cognitive radios [3].

The term “cognitive radio” was first used and discussed in [4] in 1999 by J. Mitola III and G. Q.

Maguire. In 2000, in his Ph.D. dissertation [5], J. Mitola III introduced cognitive radio and networks as

an extension of the SDR concept where the radios are computationally intelligent and are aware of the

radio environment and user applications. Although the term “cognitive radio” has been used only recently,

the concept of intelligent spectrum access is not completely new. For example, the idea of sensing before

transmitting as used in the popular carrier-sense multiple access (CSMA) protocols dates back to early

90’s when the IEEE 802.3 standards or Ethernet standards were being developed. Also, the idea of using

“dynamic channel allocation and selection” was extensively studied in the 80’s and 90’s in the context of

mobile cellular wireless networks. The fact that one of the main reasons of the spectrum scarcity problem

is the traditional fixed spectrum allocation, was first realized when the FCCs Spectrum Policy Task Force

Report [6] was published in 2002. Since then, FCC pushed towards advances in technology for more

efficient spectrum usage, and the idea of exploiting cognitive radio technology to facilitate the flexible,

efficient, and reliable use of spectrum gained momentum in both industry and academia. In particular, FCC

advocated the use of the concepts such as dynamic frequency selection (DFS), incumbent profile detection

(IPD), and transmit power control (TPC) for cognitive radios. In 2004, FCC issued an NPRM (Notice of

Proposed Rulemaking) which introduced the possibility of allowing the unlicensed (or secondary) users

to temporarily use spectrum allocated to licensed (or primary) users as long as the secondary users do

not cause harmful interference to primary users. An important milestone in the development of cognitive

radio is the approval by FCC in November 2008 of the unlicensed use of the TV white space (54-72 MHz,

76-88 MHz, 174-216 MHz, and 470-806 MHz) based on spectrum sensing as well as consultation with an

FCC-mandated database [7]. These spectrum bands are attractive due to their superior radio propagation

Page 5 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

4

characteristics in the sub-gigahertz frequencies. In September 2010, FCC released new rules for the use

of white space for unlicensed wireless devices which removed the mandatory sensing requirements and

thus facilitated the use of the spectrum with geolocation based channel allocation [8]. This has motivated

the development of new wireless standards and standardization initiatives including the IEEE 802.22 for

wireless regional area networks (WRANs), the European Computer Manufacturers Association (ECMA)

392 for personal portable devices, the IEEE 802.11af and 802.19.1 Task Groups.

In the last decade, there have been a significant amount of research work focusing on different aspects

in CRNs such as spectrum sensing and signal classification techniques at the physical layer, scheduling,

transmission power control, and adaptive channel access protocols at the link layer, resource allocation

for multi-hop transmission at the routing layer, modeling and analysis of security, and cognitive radio

network economics. In addition to a huge number of research papers including survey and tutorial type of

articles [2], [4], [9]- [14], a number of research monographs and edited books [15]- [23] have appeared

in the literature. A number of special issues on cognitive radio have been organized in both IEEE and

non-IEEE journals [24]- [36]. A number of small-scale testbeds have been built in the academia. Also,

significant amount of efforts have been put by the different standardization bodies for the standardization

of CRNs. The aim of this article is to concisely summarize the current state-of-the-art research in the

dynamic spectrum access-based cognitive radio systems and the future research trends. The rest of the

article is organized as follows. After reviewing the fundamental concepts of cognitive radio and dynamic

spectrum access, we describe the building blocks of a CRN and outline the major research issues from

a layered perspective. Then we provide a summary of the current state-of-the-art spectrum sensing and

spectrum sharing methods for cognitive radio systems. This is followed by a concise review of the work on

robustness and security in CRNs, and on the economics of CRNs. Then we describe the available cognitive

radio simulation tools, testbeds, and hardware prototypes. The current and future trends in cognitive

radio including the applications (e.g., smart grid [37] [38], machine-to-machine communications [39],

and cloud computing [40]) are reviewed and the open research issues are outlined next. To this end, the

standardization activities on cognitive radio are summarized.

II. FUNDAMENTAL PRINCIPLES OF COGNITION AND DYNAMIC SPECTRUM ACCESS

The term “cognitive radio” defined in [2] is as follows: Cognitive radio is an intelligent wireless

communication system that is aware of its ambient environment. A cognitive radio transmitter will learn

from the environment and adapt its internal states to statistical variations in the existing RF stimuli by

Page 6 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

5

adjusting the transmission parameters (e.g., frequency band, modulation mode, and transmission power)

in real-time and on-line manner. This definition essentially captures the fundamental concept behind

cognitive radio. A cognitive radio network enables us to establish communications among cognitive

radio nodes/users. The communication parameters can be adjusted according to the change in the radio

environment, topology, operating conditions, or user requirements. Two main objectives of the cognitive

radio are to improve the utilization of the frequency spectrum and to achieve the highly reliable and highly

efficient wireless communications.

Cognitive radio-based dynamic spectrum access or sharing has basically two major flavors, i.e., hor-

izontal spectrum sharing and vertical spectrum sharing. In the former case, all users/nodes have equal

regulatory status while in the latter case all users/nodes do not have equal regulatory status. In vertical

spectrum sharing, there are primary (i.e., licensed) users and secondary (i.e., unlicensed) users, and the

secondary users opportunistically access the spectrum without affecting the primary users’ performances.

Horizontal spectrum sharing can be between homogeneous networks (e.g., IEEE 802.11a operating in the

5 GHz Unlicensed National Information Infrastructure [U-NII] band) or between heterogeneous networks

(e.g., coexistence between IEEE 802.11b and 802.15.1 [Bluetooth] networks). When all the networks in

a heterogeneous environment have cognitive/adaptive capabilities (i.e., all coexisting networks have equal

incentives to adapt), it is referred to as symmetric sharing. On the other hand, when there is one or more

network without cognitive/adaptive capabilities (e.g., coexistence of legacy technology with cognitive radio

technology), this is referred to as asymmetric spectrum sharing [15]. One example of this is the coexistence

of high-speed IEEE 802.11 networks with low-power IEEE 802.15.4 networks. Dynamic spectrum access

in vertical spectrum sharing is referred to as opportunistic spectrum access. This opportunistic spectrum

access is the method for the secondary user to operate within a frequency band which is designated to

the primary user.

The major functionalities of a cognitive radio device include spectrum sensing, spectrum management,

and spectrum mobility [22]. Through spectrum sensing, the information of the target radio spectrum (e.g.,

type and current activity of licensed user) has to be obtained so that this information can be utilized

by the cognitive radio user. The spectrum sensing information is exploited by the spectrum management

function to analyze the spectrum opportunities and make decisions on spectrum access. If the status of

the target spectrum changes, the spectrum mobility function can switch the operating frequency bands for

the cognitive radio users.

• Spectrum sensing: The goal of spectrum sensing mechanism is to determine the status of the spectrum

Page 7 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

6

(e.g., to detect the signature of a signal from a licensed user) and the activity of licensed user by

periodically sensing the target frequency band. In particular, a cognitive radio transceiver detects an

unused spectrum or spectrum hole (i.e., band, location, and time) and also determines the method of

accessing it (i.e., transmission power and access duration) without interfering the transmission of a

licensed user.

Spectrum sensing can be either centralized or distributed. In centralized spectrum sensing, a sensing

controller (e.g., access point or base station) senses the target frequency band, and the information

obtained from sensing is shared with other nodes in the system. Centralized spectrum sensing can

reduce the complexity of user terminals, since all the sensing functions are performed at the sensing

controller. However, centralized spectrum sensing suffers from location diversity. For example, the

sensing controller may not be able to detect an unlicensed user at the edge of the cell. In distributed

spectrum sharing, unlicensed users perform spectrum sensing independently, and the spectrum sensing

results can be either used by individual cognitive radios (i.e., non-cooperative sensing) or can be

shared among other users (i.e., cooperative sensing). Although cooperative sensing incurs more

communication and processing overhead, the accuracy of spectrum sensing is higher than that of

non-cooperative sensing.

• Spectrum analysis: The information obtained from spectrum sensing is used to schedule and plan

the spectrum access by the unlicensed users. In this case, the communication requirements of the

unlicensed users are also used to optimize the transmission parameters. Major components in spec-

trum management mechanisms are spectrum analysis and spectrum access optimization. In spectrum

analysis, information from spectrum sensing is analyzed to understand the ambient RF environment

(e.g., the behavior of licensed users) and gain knowledge about the spectrum holes (e.g., interference

estimation, duration of availability, and probability of collision with licensed user due to sensing

error). A knowledge base of the spectrum access environment can be built and maintained based on

learning and knowledge extraction. Machine learning algorithms from the field of artificial intelligence

can be applied for learning and knowledge extraction. Subsequently, a decision to access the spectrum

(e.g., frequency, bandwidth, modulation mode, transmission power, location, and time duration) is

made by optimizing the system performance given the desired objective (e.g., maximize throughput

of the unlicensed user) and constraints (e.g., maintain the interference caused to the licensed users

below the target threshold).

• Spectrum access: After the decision is made on spectrum access based on spectrum analysis, the

Page 8 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

7

spectrum holes (also called spectrum opportunities) are accessed by the unlicensed users. Spectrum

access is performed based on a cognitive medium access control (MAC) protocol which intends to

avoid collision/harmful interference with/to licensed users and also with other unlicensed users. The

cognitive radio transmitter is also required to perform negotiation with the cognitive radio receiver

to synchronize the transmission so that the transmitted data can be received successfully. A cognitive

MAC protocol could be based on a fixed allocation MAC (e.g., frequency-division multiple access

[FDMA], time-division multiple access [TDMA], code-division multiple access [CDMA]) or a random

access MAC (e.g., ALOHA, carrier-sense multiple access with collision avoidance [CSMA/CA]).

The optimal spectrum access decision depends on the ambient environment and the cooperative or

competitive behavior of the unlicensed users. The spectrum access decision can be made in a non-

cooperative and distributed way based on a local optimization objective. Alternatively, a cooperative

spectrum access decision can be made either in a distributed or a centralized way based on a global

optimization objective. The spectrum access decisions are then communicated among cognitive radio

transmitters and receivers.

In a cognitive radio network, the secondary users may use either an interference control (or spectrum

underlay) or an interference avoidance (or spectrum overlay) approach to exploit the spectrum

opportunities. In the spectrum underlay approach, the secondary users transmit over the same spectrum

as the primary users as long as the interference caused to the primary users does not exceed a threshold

level. Therefore, such an approach requires a sophisticated power control scheme for secondary

transmitters. In the spectrum overlay approach, the secondary users need to have the knowledge about

spectrum holes so that the secondary users can exploit them ensuring that there is no interference

caused to the primary users. The interference avoidance approach is, therefore, more conservative

than the interference control approach, and no strict power control is required for this spectrum access

paradigm.

• Spectrum mobility: Spectrum mobility is a function related to the change of operating frequency band

of cognitive radio users. When a licensed user starts accessing a radio channel which is currently

being used by an unlicensed user, the unlicensed user can switch to the idle spectrum band. This

change in operating frequency band is referred to as spectrum handoff [9]. During spectrum handoff,

the protocol parameters at the different layers in the protocol stacks have to be adjusted to match

with the new operating frequency band. Spectrum handoff must ensure that the data transmission by

the unlicensed user can continue on the new spectrum band.

Page 9 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

8

The above functionalities can be implemented based on the software-defined radio (SDR) which is a

reconfigurable wireless communication system. With SDR, the transmission parameters (e.g., operating

frequency band, modulation mode, and protocol) can be controlled dynamically. This adjustability function

is achieved by software-controlled signal processing algorithms. The main functions of SDR include multi-

band operation, multi-standard support, multi-service support, multi-channel support, and self-organization

(i.e., “plug-and-play” capability). The flexibility of SDR is useful for prototyping and evaluating cognitive

networking technology.

In a network of cognitive radios, the network infrastructure should support and facilitate the above

cognitive radio functionalities (i.e., spectrum sensing, analysis, access, and mobility). The network will

facilitate the communication among multiple cooperating cognitive radios, or the network will control

communication processes among competing cognitive radio users through proper radio resource allocation.

The policies specifying the limitations of radio access (e.g., available frequency bands in a certain location)

and the operating characteristics of the physical radio can be stored in the policy and configuration

databases which would be accessible to the cognitive radios through the network infrastructure.

III. ARCHITECTURES AND BUILDING BLOCKS OF COGNITIVE RADIO NETWORKS: A LAYERED

PERSPECTIVE

A dynamic spectrum access-based cognitive radio network may adopt either a cooperative or a non-

cooperative network architecture. In the non-cooperative CRN architecture, each cognitive node is respon-

sible for its own decision. Therefore, it has minimal communication requirements (hence less overhead);

however, the spectrum utilization may be low. On the other hand, in the cooperative but centralized CRN

architecture, a centralized server maintains a database of spectrum availability and access information

(based on information received from secondary users, e.g., through a dedicated control channel). Therefore,

spectrum management is simpler and coordinated in this case, and it enables efficient spectrum sharing.

For dynamic spectrum access, a cooperative but distributed CRN architecture relies on cooperative local

actions throughout the network to achieve a performance close to the global optimal performance. In this

architecture, cognitive radios periodically exchange information on their local environment, communication

requirements and performances among themselves. The cognitive radios use their local information as

well as information received from their peers to determine their communication parameters. The network

performance, however, may suffer from the hidden node problem and large control overhead. Note that

in both centralized and distributed strategies, the primary user may or may not cooperate. The protocol

Page 10 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

9

architecture for a cognitive radio system is shown in Fig. 1. The communication protocols at the different

layers need to function such that the utilization of the radio spectrum is maximized while satisfying the

policy constraints.

Physical layer (PHY): In the physical layer (PHY), spectrum sensing is the primary task which includes

functionalities such as the following: detect spectrum holes/opportunities over wide frequency spectrum,

opportunity estimation, estimate interference at the primary receiver (e.g., in a CDMA environment).

Spectrum sensing involves obtaining the spectrum usage characteristics across multiple dimensions such

as time, space, frequency, and code (e.g., modulation, waveform, bandwidth, carrier frequency, etc). A

primary transmitter can be detected by using methods such as matched filter detection, energy detection,

and cyclostationary feature detection. The leakage power at the RF front-end of a primary receiver can

be measured/estimated to detect the activity of the corresponding primary receiver. In the physical layer,

the RF front-end is implemented based on software-defined radio (i.e., SDR transceiver). The RF front-

end requires high sampling rate, high resolution analog-to-digital converters (ADCs) with large dynamic

ranges, multiple analog front end circuitries, and high-speed signal processors to perform computationally

demanding signal processing tasks with small delay.

MAC layer: In the MAC layer, a decision is taken on whether to transmit or not considering that

the spectrum sensing/detection could be erroneous and how to exploit the spectrum holes (e.g., what

modulation and power level to use), and in case of transmission how to share the spectrum with other

cognitive radios. The MAC protocol functionalities include the following: obtain information on channel

occupancy (spectrum sensing1) and make decision on spectrum access; synchronize transmission param-

eters (e.g., channel and time slot) between transmitter and receiver; facilitate negotiation among primary

users and secondary users for spectrum allocation; facilitate communication among secondary users to

perform channel sensing and channel access; and facilitate spectrum trading functions (e.g., spectrum

bidding and spectrum pricing). The major challenges in designing the MAC protocols for CRNs include

optimal channel sensing for multi-channel access; primary users’ time-varying activity; hidden and exposed

terminal problems; synchronization between transmitter and receiver for which the unavailability of any

fixed common control channel and the channel availability changes depending on the spectrum access by

primary users; optimal channel allocation/scheduling, rate and power adaptation; and coexistence between

primary and secondary users.

1The spectrum sensing in the physical layer relates to signal detection and processing, while the spectrum sensing in the MAC layerinvolves the scheduling of a spectrum sensing activity.

Page 11 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

10

Network layer: The primary tasks at the network layer are topology construction, addressing, and routing.

The topology construction involves spectrum detection, neighbor discovery, and topology management

(e.g., through spectrum mobility). The addressing can be fixed (e.g., extension of PHY/MAC address) or

dynamic (e.g., using dynamic host configuration [DHCP] protocol). In multi-hop CRNs, routing decisions

need to be made based on topology, MAC congestion, and link quality/reliability (a cross-layer solution).

Note that there is no fixed common control channel (e.g., for route discovery) and rerouting may need to

be performed frequently due to intermittent connectivity among cognitive nodes. The routing metric has

to consider parameters such as hop-count and link quality as well as spectrum management parameters

such as interference to primary users and availability of spectrum holes.

Transport layer: The transport layer protocol is responsible for flow and congestion control, the per-

formance of which is affected by MAC protocol performance and spectrum mobility. The throughput

performance of the traditional transmission control protocol (TCP) is a function of round-trip time

(RTT) and packet loss probability, which would depend on the spectrum management/sharing protocols,

bandwidth of spectrum holes, transmission power and interference levels. During rerouting, RTT and

packet loss rate change, and spectrum handoff latency (due to spectrum mobility) may increase the value

of RTT. Therefore, the performance of TCP in a CRN would be improved if the packet delay and loss

can be minimized during spectrum handoff (e.g., through efficient queue management).

The adaptive protocols in the MAC, network, transport, and application layers should be aware of the

variations in the cognitive radio environment. In particular, the adaptive protocols should consider traffic

activity of primary users, transmission requirements of secondary users, and variations in channel quality.

For this, the MAC-layer beacons can be sensed for information about sleep and wakeup schedules of the

cognitive radios and information about nodes’ connectivity. The routing protocol messages, for example,

can be exploited to detect mobility of the cognitive radio nodes. To link all the protocol modules in the

stack and enable cross-layer interaction, a cognitive radio control is used to establish the interfaces among

SDR transceiver, adaptive protocols, and wireless applications and services. This cognitive radio module

(Fig. 1) uses intelligent algorithms to process the measured signal from the physical layer, and to receive

information on transmission requirements from the applications to control the protocols parameters in the

different layers.

A programmable radio architecture and related protocol stack with cross-layer interactions as described

above along with the wideband radio front-ends will be the key enabling technologies for the cognitive

radio networks. In the following section, we survey the existing literature on PHY, MAC, and networking

Page 12 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

11

Applications and services

Transmit/ receive

Cogn

itive

radi

o

SDR transceiver

MAC

Network

Transport

Adap

tive

prot

ocol

Fig. 1: Cognitive radio protocol stack.

layer issues in cognitive radio networks.

IV. CURRENT STATE-OF-THE-ART RESEARCH ON SPECTRUM SHARING IN COGNITIVE RADIO

NETWORKS

In this section, we provide a survey on the evolution of cognitive radio research covering different

aspects of spectrum sharing, i.e., spectrum sensing [41]- [61], measurements and statistical modeling of

spectrum usage [62]- [68]; physical layer (PHY) aspects such as waveform and modulation design [69]-

[92]; multiple access, resource allocation and power control, and spectrum mobility [93]- [125]; cognitive

learning, adaptation and self-configuration [126]- [140], and multi-hop transmission and routing [141]-

[148].

A. Spectrum Sensing, Interference Modeling, Measurements and Statistical Modeling of Spectrum Usage

Spectrum statistics are important information for secondary users to opportunistically access spectrum

of primary users. Therefore, there is a need to study and understand the characteristics and natures of

spectrum usage by the primary users. To gain such information, the following issues are considered in

the literature.

• Spectrum sensing: Spectrum sensing is a primitive action for secondary users to detect the status of

spectrum access by primary users. Without this information, the secondary users may not access idle

Page 13 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

12

spectrum, reducing utilization, or may cause interference to the primary users occupying spectrum.

• Interference modeling: Secondary users may observe the interference on spectrum for two reasons.

Firstly, the secondary users must ensure that their transmission will not interfere and interrupt the

ongoing transmission of primary users. Secondly, given an interference condition, the secondary users

must access spectrum such that their transmission requirements are satisfied. Interference modeling

provides the secondary users with the capability to achieve these goals.

• Measurements and statistical modeling of spectrum usage: While spectrum sensing is a short-term

action to observe an instantaneous status of spectrum, spectrum measurement is performed in a long-

term basis (e.g., over a few months) to gain knowledge and statistical data of primary users. This

information is useful for the secondary users to determine their spectrum access strategy (e.g., access

in a particular time of the day to minimize interference to the primary users).

In the following, we provide reviews for the works related to these issues.

An overview of the regulatory requirements (e.g., sensing periodicity and detection sensitivity) and

the major challenges associated with spectrum sensing in cognitive radio networks are given in [41],

[42]. A comprehensive survey of the different spectrum sensing techniques can be found in [43], [44].

Traditional spectrum sensing techniques include the following: matched filter detection, energy detection,

cyclostationary detection, wavelet detection, and covariance detection. The performance metrics generally

used for spectrum sensing are: probability of correct detection and probability of false alarm. The main

challenges of spectrum sensing arise due to the uncertainty in a radio environment caused by channel

fading, interference, and noise, requirement of quick detection, and also due to the fact that the modulation

schemes, transmission powers, bit rate may be different for the different primary users. Also, a secondary

user may not be able to detect the location of the primary receiver. In this context, we refer to [46] for a

comprehensive overview of the different channel fading models. In [47], the authors evaluate the secondary

channel capacities under the average and peak interference power constraints at the primary receiver in

different fading environments such as Rayleigh fading, Nakagami fading, and log-normal shadowing. It

is shown that with the same interference power limit, channel capacity in fading environments exceeds

that of the AWGN channel.

In [48], the authors use random matrix theory to obtain the probability distributions of the test statistics

and find the closed-form expressions for the probability of detection and the probability of false alarm.

Also, the authors propose spectrum sensing methods which overcome the noise uncertainty problem, and

can be used without requiring the knowledge of signal, channel, and noise power. [49] studies the impact of

Page 14 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

13

the uncertain information about the noise power level on the performance of the energy detector spectrum

sensing. The analysis of an energy detector with estimated noise power (ENP) is presented. This analysis

can be used to determine the threshold and establish conditions for existence of the SNR wall. In [51],

the authors provide a tutorial on the multi-taper method (MTM), which is a nonparametric method for

spectrum sensing. This method provides high spectral-resolution capability, estimates the average power

in each sub-band of the spectrum, and identifies the unknown directions of interfering signals. In [52], the

spectrum sensing duration is optimized to maximize the throughput of secondary users while achieving a

target detection probability in order to protect the primary users.

For improved performance, cooperative spectrum sensing can be used [53]- [60]. The cooperative

spectrum sensing process works as follows. Every cognitive radio performs its own local spectrum sensing

independently and then makes a binary decision on the presence of a primary user. Then, all of the CRs

transmit their decisions to a common node. The common node fuses the decisions and makes a final

decision to identify whether a primary user is present or not. The fusion techniques of cooperative spectrum

sensing can be either decision fusion or data fusion. Through cooperation, the reliability of primary user

detection can be improved. However, the performance of cooperative sensing can be limited when the

reporting channels are subject to fading and/or shadowing. In [53], the authors propose several robust

cooperative spectrum sensing techniques based on the concepts of cooperative diversity and multiuser

diversity.

In [56], the authors focus on the cooperative spectrum sensing system in which the individual cognitive

radio users make independent decisions about the presence of the primary signal and communicate their

decisions to a fusion center. Then, the fusion center makes the final decision about the occupancy of the

band. The observations at the nodes are assumed to be correlated, and therefore, a linear-quadratic (LQ)

fusion strategy is proposed as compared with counting rule which is linear combining. The proposed LQ

rule requires the second-order statistics of the local decisions when the primary users are present, which

can be obtained by estimating the pairwise statistics. In [58], the authors present optimal weightings for

linearly combining the energies measured at the cognitive radio users such that the probability of detection

is maximized with a constraint on the probability of false alarm. The amount of sensing and transmission

overhead in a cognitive radio network with large number of cooperating users and frequency channels

is large when every node scans all the channels and sends each channel’s status to the fusion center. To

reduce the communication overhead, [59] proposes matrix completion and joint sparsity algorithms for

spectrum sensing based on a small number of observations.

Page 15 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

14

The work in [60] proposes a location-aware cooperative sensing algorithm which decides if a primary

user exists within a certain detection area or not. The primary network is modeled as a random geometric

network and the cognitive radio network is assumed to know the relative locations of its users. The optimal

weightings, that linearly combine the measured energy levels of the sensing nodes such that a modified

deflection coefficient is maximized, are obtained by Fisher linear discriminant analysis.

The work in [62] reports a comprehensive experimental study on interference analysis in the IEEE

802.22-based wireless regional area network (WRAN) system. The work in [42] is a significant contribution

towards modeling the distribution of aggregate interference at a primary receiver due to the secondary

transmitters in terms of system parameters of a spectrum sensing-based cognitive radio network. The

work in [63] is also an important contribution to statistical modeling of aggregate interference caused to

a primary user in a cognitive radio network. The theory of truncated stable distributions is used for the

modeling. The effect of power control on the cognitive network interference is also considered. [64] is

one of the very early studies on spectrum occupancy and interference in cognitive radio systems. [64]

discusses the opportunities, challenges, and communication limits in the cognitive radio technology. [65]

provides a framework for spectrum measurement and offline processing of measured data. The framework

has been evaluated by using on real-world spectrum measurement data. The authors in [66] propose a

statistical model for spectrum occupancy in time and frequency domain, and the key model parameters

are determined from actual measurements. The work in [67] presents a spectrum measurement campaign

in detail and the lessons learned in this campaign. Also, the paper presents a stochastic model based on

modified beta distribution for the duty cycle of occupancy in a sub-band. The authors in [68] address the

potential drawback of Poisson modeling for primary user activity and introduces a new model based on

the first-difference filter clustering and temporal correlation statistics.

There are a few design considerations among different spectrum sensing methods. Firstly, the cooperative

spectrum sensing can improve the detection performance, but at the cost of higher false alarm probability

and larger overhead (e.g., processing and communication). Spectrum sensing can be performed in a reactive

or proactive manner. A reactive spectrum sensing method will not incur an overhead if spectrum access is

not performed as in a proactive spectrum sensing. However, the reactive spectrum sensing will have longer

latency to collect radio environment status when a secondary user needs to access spectrum. Finally, the

tradeoff between rate and reliability must be optimized. Spending longer time in spectrum sensing will

lead to a more accurate sensing result. However, this is at the cost of lower rate due to shorter time for

transmission.

Page 16 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

15

B. Waveform and Modulation Design for Cognitive Radios

To minimize interference to primary users, the design of waveform and modulation of signal from

secondary users can be optimized. For example, in an underlay spectrum access scheme, the secondary

user can transmit based on ultra-wideband (UWB) and adjust the pulse position and/or width to avoid

interference to the narrow band transmission of the primary users. Similarly, in an overlay spectrum access

scheme, the secondary user adapts the multicarrier modulation for an OFDM-based system to minimize

interference.

1) Waveform design and interference mitigation: In [69], a new pulse waveform design method is

proposed for the Ultra Wideband (UWB) telecommunication system which has to coexist with some

other narrowband systems. In [72], a novel low-complexity adaptive UWB pulse shaping algorithm is

presented for producing the expected spectral notches right in the frequency band occupied by the nearby

wireless devices. Adaptive pulse waveform optimization based on linear combination of different derivative

functions of Gaussian pulse is investigated to produce expected notches in achieving narrowband inter-

ference avoidance as well as matching with the FCC spectral mask. In [75], non-contiguous multicarrier

modulation that can be used to facilitate spectrally opportunistic cognitive radio, is considered. [76] jointly

designs the transmitted waveform and the FIR filters at the relay nodes for cognitive relay networks. [77]

considers the problem of waveform optimization for multiple input single output cognitive radio links with

the spectral mask constraint at the transmitter and the interference cancellation at the receiver. In [86],

a novel method of adding extended active interference cancellation signals to suppress sidelobes and to

shape the spectrum of CR-OFDM signals with a cyclic prefix (EAIC-CP) is presented. The key idea of

the proposed EAIC-CP method is to use the cancellation signals consisting of EAIC tones that are spaced

closer than the OFDM subcarrier interval.

In [87], a statistical model of interference aggregation in spectrum sensing-based cognitive wireless

networks is presented by explicitly taking into account the random variations in the number, location

and transmitted power of the cognitive radios as well as the propagation characteristics. [90] proposes

an enhanced opportunistic interference mitigation scheme utilizing both successfully and unsuccessfully

decoded primary packets to improve the data rate of secondary user’s transmission. Also, an analytical

model is proposed to investigate the overall spectral efficiency due to the changes of spectrum usage

affected by the coexisting secondary user’s transmission.

2) Modulation design and classification: [91] provides an excellent tutorial on two multicarrier mod-

ulation schemes, namely, non-contiguous orthogonal frequency-division multiplexing (NC-OFDM) and

Page 17 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

16

non-contiguous non-orthogonal frequency-division multiplexing (NC-NOFDM) for opportunistic wireless

access. [78] also reviews different modulation schemes (e.g., DFT-spread OFDM, constant envelope OFDM

and filter bank multicarrier) for cognitive radios with their advantages and disadvantages. Different mod-

ulation schemes are compared in terms of spectral characteristics, signal metrics, and system performance

with the presence of non-linear distortions. In [84], a quantitative comparison of NC-OFDM and non-

contiguous multicarrier CDMA transmission techniques is presented in the context of a dynamic spectrum

access network. In [85], the authors provide a review of the state-of-the-art methods specifically designed

to protect the transmissions of the primary users from possible interference caused by nearby secondary

transmitters using NC-OFDM. These methods are designed to suppress out-of-band (OOB) emissions

resulting from the use of NC-OFDM transmission.

In [79], the performance of adaptive modulation for cognitive radio with opportunistic access is analyzed

by considering the effects of spectrum sensing and primary user traffic for Nakagami-m fading channels.

Both the adaptive continuous-rate scheme and the adaptive discrete-rate scheme are considered. [83]

introduces modulation strategies to realize the coexistence between the CR-based rental system and the

licensed system in such a way that the secondary users are invisible to the primary users.

In [89], the techniques for detecting interfering signals, estimating the distance between victim and

interfering devices, and shaping the spectrum to mitigate interference into the victim receiver are con-

sidered. [80] designs a classifier for linear digital modulation based on the clustering (using the K-

Means algorithm) of the received signal constellation diagram. [82] presents a novel digital modulation

classification system for cognitive radios using only temporal waveform features. A hierarchical approach

is used to first make separations into intermediate subclasses, where some of the subclasses can consist

of more than one modulation type. Then the second classifier is used to discriminate between higher

order modulation schemes using additional features. In [92], the authors propose a method for modulation

classification based on the Support Vector Machines (SVMs) which is robust against heavy non-Gaussian

interference. In the experiment, four classes of signal modulations, i.e., binary phase-shift keying (BPSK),

quadrature phase-shift keying (QPSK), frequency-shift keying (FSK), and minimum shift keying (MSK),

are considered.

C. Analysis of Spectrum Sharing Systems

Capacity analysis is still an important issue to quantify the upper bound performance of the cognitive

radio system. The capacity of a cogntiitve radio system is constrained not only by the transmit power

Page 18 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

17

limit, but also by the interference to primary users. This analysis is useful to determines the achievable

transmission rates of secondary users. The analysis can consider the effects of different channel fading.

[47] evaluates the secondary channel capacities under the average and peak interference power con-

straints at the primary receiver in different fading environments. It is shown that with the same interference

power limit, channel capacity in fading environments exceeds that of the AWGN channel. The impacts

of correlated fading and multiple primary receivers on the channel capacity are also studied. Asymptotic

capacities under different fading distributions are derived. The work in [93] analyzes the ergodic capacity of

secondary user system for different spectrum sharing strategies (underlay, overlay, and hybrid strategies)

with one primary user and one secondary user. The analysis reveals that the maximum capacity in a

spectrum underlay system can be achieved with reduced signaling complexity where the channel state

information between the secondary transmitter and the primary receiver may not be required for power

allocation.

In [94], the authors discover a new way of spectrum sharing by exploiting the primary Automatic

Repeat and reQuest (ARQ) retransmissions for secondary user communications. The paper also avoids

somewhat impractical assumptions such as complete CSI at the secondary transmitter. [110] presents a

learning-based spectrum sharing protocol for spectrum-sharing wireless networks such that the weighted

time-fairness can be achieved among the different networks sharing the spectrum.

D. Multiple access, Resource Allocation, and Power control

1) Multiple access in spectrum underlay: In a spectrum underlay scenario, the problem of optimal

spectrum sharing among secondary users can be formulated as an optimization problem with a suit-

able objective function and a set of constraints which capture user fairness, quality-of-service (QoS) of

secondary users, and interference constraints for primary users as follows:

max U(R1, R2, . . . , Rn)

subject to

Ri ≥ Bi, i = 1, 2, . . . , n

µj =n∑

i=1

hj,iPi ≤ Ij, j = 1, 2, . . . ,m

where U is the objective function, Ri is the achievable rate for secondary user i, Bi is the minimum

rate requirement for user i, µj is the total interference created by secondary users at primary receiver j,

Page 19 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

18

hij denotes the channel gain from the transmitter of secondary user i to primary receiver j, Pi denotes

transmission power of secondary transmitter i, and Ij denotes the maximum tolerable interference level

at primary receiver j.

Note that the optimization problem above may not be feasible when its constraints are too stringent

and/or the network load is too high. If this is the case, an admission control mechanism needs to be

invoked to limit the number of admitted secondary users. Then, the power allocation for the set of admitted

secondary users can be performed. Using this framework, in [96], a solution approach is proposed for

the joint admission control and power allocation problem for secondary users to achieve fairness among

them assuming CDMA technology at the physical layer. A related spectrum sharing problem, which

maximizes throughput of secondary networks, is solved in [97]. To obtain the power allocation solutions,

the instantaneous channel gains among secondary users and interference from secondary transmitters

to primary receivers need to be estimated. In a practical scenario where only estimates of the average

channel gains are available, the power allocations for the secondary transmitters need to be done in a

conservative way to satisfy the target interference constraint violation probability for primary receivers [98].

The problem of sum-rate maximization for secondary transmitters under joint beamforming and power

allocation in a cognitive radio network with multiple secondary transmitters and primary receivers (each

with one antenna) is investigated in [100].

The performance of a cognitive radio network could be enhanced by using multiple antennas at the

secondary transmitters and/or secondary receivers. However, in this case, power allocations among the

transmit antennas under total transmission power constraints and interference-power constraints at the

receiver antennas need to be jointly considered. Also, the tradeoff between throughput and fairness in a

multiuser MIMO cognitive radio network needs to be modeled and analyzed. The work in [99] exploits

multiple antennas at the secondary transmitter to maximize the cognitive radio’s transmission rate under

the transmission power constraint and a set of interference power constraints. The authors demonstrate the

effectiveness of using convex optimization for solving resource allocation problems for cognitive spectrum

sharing.

A survey on dynamic resource allocation schemes for cognitive radio systems with the interference

temperature based spectrum sharing model can be found in [101]. To maximize the secondary network

throughput, the transmission power, bit-rate, bandwidth, and antenna beam can be dynamically allocated

based upon the available channel state information (CSI) of the primary and secondary networks. Many

new and challenging problems regarding the design of CR systems are formulated, and some of the

Page 20 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

19

corresponding solutions are shown to be obtainable by restructuring some classic results known for

traditional (non-CR) wireless networks.

2) Multiple access in spectrum overlay: In a spectrum overlay scenario, secondary users need to sense

a frequency band of interest and transmit only if primary users are not detected on the chosen band.

Therefore, no strict power control is required for this spectrum access paradigm. Either scheduling or

random access-based multiple access techniques can be used [102], [103]. In the former case, a scheduler

(e.g., in a base station or an access point) makes the scheduling decisions based on the spectrum sensing

results and/or the statistical information regarding channel availability. Generally, either a Markovian

or an independence assumption is made for channel availability. In the latter case, the secondary users

need to discover the spectrum opportunities to access the channel(s) through a contention resolution and

reservation process. The hardware constraints (i.e., partial spectrum sensing and spectrum aggregation

limit) of secondary users need to be considered for a practical channel access protocol.

The work in [104] studies the problem of optimal discovery of spectrum opportunities with MAC-

layer sensing in cognitive radio networks. The objectives are to maximize the discovery of spectrum

opportunities by adapting sensing periods and to minimize the delay in finding an available channel. The

analysis defines sensing overhead and unexplored opportunity, and optimizes these factors. The sensing

overhead is the average fraction of time during which a channel’s discovered opportunities are interrupted

and not utilized due to sensing of one of the channels. The unexplored opportunity is the average fraction of

time during which channel’s opportunities are not discovered in case that the channel is being periodically

sensed with its sensing period. The optimal channel sequencing can be obtained from the probability that

the channel would be idle at a certain time based on the previous samples. The proposed approach requires

the estimation of the underlying channel-usage patterns with an ON/OFF alternating renewal channel. The

distribution parameters of ON/OFF periods are estimated by Maximum Likelihood (ML) estimators and

their confidence intervals.

With a Markovian assumption on channel availability, the work in [105] proposes decentralized cognitive

medium access method in a single-user scenario which allows a secondary user to independently search

for spectrum opportunities without a central coordinator or a dedicated communication channel. An

analytical framework is developed based on the theory of Partially Observable Markov Decision Process

(POMDP). However, the optimal policy for a general POMDP can be computationally prohibitive due to

the exponential growth of the dimension of the sufficient statistic. The authors propose a greedy approach

that maximizes per slot throughput. This decision-theoretic approach integrates the design of spectrum

Page 21 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

20

access with spectrum sensing at the physical layer and traffic statistics determined by the application layer

of the primary network. An opportunistic user makes optimal decisions for sensing and access based on

the belief vector that summarizes the knowledge of the network state based on all past decisions and

observations.

With an independence assumption on channel availability, [106] proposes a hardware-constrained mul-

tichannel cognitive MAC protocol considering the tradeoff between exploring more idle channels and

encountering more sensing overhead. The channel access problem is formulated as a well-defined optimal

stopping problem. The objective is to choose a time to stop spectrum sensing (probing) such that the

expected reward is maximized. The decision is either to stop and receive the known reward or to continue

and observe the channel status (i.e., the 0-1 (occupied/idle) state of the channel probed) for further decision.

This problem can be solved by the backward induction method. The similar case of correlated traffic from

primary users is also considered in [107].

3) Cross-layer multiple access: In [108], the authors characterize quantitatively the interaction between

the physical and MAC layers. In particular, they demonstrate how the sensing errors at the PHY layer affect

the MAC design and how incorporating the MAC layer information into physical layer leads to a cognitive

spectrum sensor whose performance improves over time by learning from accumulating observations.

The cross-layer channel access problem is formulated as a constrained POMDP. It is shown that there

exists a separation principle in the optimal joint design of channel access that circumvents the curse of

dimensionality in general POMDPs and leads to closed-form optimal solutions.

[111] proposes a cross-layer based opportunistic channel access protocol that integrates spectrum

sensing at the physical layer and packet scheduling at the MAC layer. The authors assume that each

secondary user is equipped with two transceivers. One transceiver is tuned to the dedicated control channel

and another transceiver is used to periodically sense and dynamically use the identified unused channels.

Two channel spectrum-sensing policies, namely, the random sensing policy and the negotiation-based

sensing policy, are considered. Under the random sensing policy, each secondary user randomly selects

one of the channels for sensing with a uniform distribution. Under the negotiation-based sensing policy,

each secondary user attempts to select a channel for sensing by overhearing the control packets in the

control channel. The authors develop an M/GY /1-based queuing model to analyze the performance of

the proposed multi-channel MAC protocols under the two types of channel-sensing policies for both the

saturation and non-saturation conditions. The analyses reveal the tradeoff between throughput and delay,

which would be useful to support the different QoS requirements in cognitive radio networks.

Page 22 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

21

In multichannel cognitive radio networks, one important issue is how a secondary user chooses the

sequence of channel sensing such that the best idle channel can be identified within the shortest time

period, maximizing throughput of the secondary user. [112] considers the optimal channel sensing order

by formulating and solving the dynamic programming model. This optimal sensing order is part of a

cognitive medium access control (MAC) which can be implemented given a spectrum sensing technique

in a physical layer. In the simplest case, [112] first shows that the intuitive sensing order (i.e., descending

order of the channel availability probabilities) is optimal. Then, the model is extended for the system with

adaptive modulation transmission, when the probability of an idle channel is known. Finally, the most

complex scenario, where the probability of an idle channel is unknown and there is a sensing error, is

also considered.

Cross layer design and optimization can be applied for multiple access of cognitive radio nodes. In

this case, the MAC protocol can utilize information from protocols in other layers (e.g., physical and

network layers) to optimize its performance. However, the cross layer approach may suffer from losing

independency of a protocol in each layer. That is, an operation of a protocol in one layer depends on a state

of protocols in other layers. Nevertheless, a few works study the benefits of cross layer design in cognitive

radio networks. [113] integrates spectrum sensing in the physical layer with the MAC protocol. Specifically,

the collaborative spectrum sensing policies are introduced, i.e., random sensing and negotiation-based.

While the former is a simple policy where secondary users can decide on spectrum sensing and access

randomly and independently, the latter allows the coordination so that the secondary users choose different

channels to sense and access. [114] studies the use of location information of a primary user to optimize

the PHY-MAC protocol of secondary users. The location information is obtained through observing RSSI

and direction of signal from an antenna array. The collaborative sensing algorithm is proposed to utilize

this location information and optimize the spectrum access given the interference constraint to the primary

user. The similar problem is also studied in [115] with the reporting overhead taken into account.

4) Power control: In [116], the authors discuss the coexistence of the primary and cognitive users in the

same spectrum by using power allocation strategies at the cognitive radios such that the interference power

experienced by the primary receiver is limited. The work considers various combination of peak/average

transmission and interference power constraints, and studies the power allocation strategies to achieve

the ergodic, delay-limited, and outage capacities of the cognitive radio system. Different fading channel

models such as Rayleigh, Nakagami, and log-normal shadowing are considered.

The orthogonal frequency division multiplexing (OFDM)-based transmission technology is particularly

Page 23 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

22

promising for cognitive radio networks using spectrum overlay. However, OFDM transmission causes

mutual interference between the primary and secondary users due to the non-orthogonality of the trans-

mitted signals. The amount of interference depends on the power allocated in the subcarriers as well

as the spectral distance between subcarriers used by the primary and secondary users. In [117], the

authors investigate the optimal power loading problem to maximize the transmission data rate while

maintaining the interference caused to the primary users within a given limit. The work is extended

in [118] considering different statistical interference constraints imposed by different primary receivers

where the probability that the interference introduced by secondary user’s transmission in the ith subcarrier

to the lth primary receiver (∑N

i=1 I(l)i (dil, Pi)) is less than the interference limit (I(l)th ) must be greater than

a (i.e., Prob(∑N

i=1 I(l)i (dil, Pi) ≤ I

(l)th

)≥ a).

[120] considers the problem of energy-efficient power allocation for maximizing the expected trans-

mission rate for OFDM-based cognitive radio systems. The reliability of the available sub-bands (which

depends on sensing error and primary user’s activity), sub-band power constraints, and total allowed

interference limit to the adjacent primary user bands are taken into account. The authors consider an

energy-aware capacity expression through the factor called subcarrier availability. A convex optimization

problem is formulated which incorporates a linear average rate loss function in the optimization objective

to include the effect of subcarrier availability. The objective function is to maximize the capacity minus

cost due to transmit power.

Due to the complex structure of the optimal solution, the authors propose three suboptimal schemes,

namely, the step-ladder, nulling, and scaling schemes, to redistribute the power within the sub-bands

which share boundaries with the primary users’ bands. The step-ladder scheme assigns lower power to

subcarriers closer to primary users’ bands with a step size which could be either constant or proportional

to the interference caused to a primary user. For the nulling mechanism, zero power is allocated to

subcarriers adjacent to the primary user band (one-nulling) or zero power is allocated to two subcarriers

closest to primary user band on each side (two-nulling). Finally, the scaling scheme allocates the power

that satisfies the interference criterion strictly by scaling down the power to the function on the total

interference threshold prescribed by primary user bands. The worst-case complexity of the algorithm is

O(MN), where M is the number of groups of the available sub-bands and N is the available subcarriers

to the secondary user for opportunistic access using OFDM.

[121] extends the traditional scheduling and power control problems by considering their robustness for

cognitive spatial-reuse time division multiple access (STDMA) network. The objective is to minimize the

Page 24 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

23

transmission length (in terms of time slots) of the secondary links under their minimum QoS requirements

without violating the maximum tolerable interference limit for the primary receivers. Due to the stochastic

nature of wireless link gains, the power allocation based on average (or estimated) link gains can be

improper. To overcome the improper power allocation, the channel gain uncertainty is modeled by ellipsoid.

Since an optimal solution cannot be obtained due to the NP-completeness of the problem, the authors

propose a distributed two-stage algorithm based on the distributed column generation method to obtain

the near-optimal solution for the robust transmission schedules in an ad-hoc cognitive radio network.

E. Distributed Learning, Adaptation, Self-configuration, and Coexistence

Learning, adaptation, and self-configuration are the fundamental functions of cognitive radio. Secondary

users should be able to observe spectrum state and learn the status of a primary user, then optimize and

adapt the transmission so that the secondary users can achieve their performance target while minimizing

interference to the primary user. Self-configuration is also an important functionality for the secondary

users to determine a set of optimal system parameters. The self-configuration capability generally relates

to a long-term setting (e.g., frequency band and transmission range), while the dynamic spectrum access is

concerned about a short-term action (e.g., transmit power and time slot to access). The main aim of which

is to ensure that a secondary user can coexist with primary users and other secondary users amicably.

The work in [126] describes the motivations, architecture, functionality, design and implementation of

cognitive networks applicable to both wired and wireless networks. A cognitive process to learn from past

decisions and use this learning to influence future behavior is the foundation for such networks.

In [127], the authors propose a game theoretic framework to analyze the behavior of cognitive radios

for distributed adaptive channel allocation. The objective is to choose the frequency channels which

maximize the reward function. Two different reward functions are formulated to capture the utility of

selfish users (U1) and cooperative users (U2), respectively. U1 values a channel based on the level of

interference perceived on that particular channel while U2 accounts for the interference seen by a user on

a particular channel, as well as the interference this particular choice will create to neighbouring nodes.

The authors demonstrate that the channel allocation problem can be formulated as a potential game and

thus it converges to a deterministic channel allocation Nash equilibrium point. Although the game with

the utility function given by U2 fits the framework of an exact potential game, the function U1 lacks

the necessary symmetry properties that will ensure the existence of a potential function. To analyze the

behavior of the selfish user game, the authors resort to the implementation of adaptation protocols using

Page 25 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

24

no-external-regret learning algorithm with exponential updates as described below.

By the proposed learning adaptation process, the CR users learn how to choose the channels that max-

imize their rewards through repeatedly playing of the game. The no-regret learning provides performance

similar to that of the potential game when cooperation is enforced, but with a higher variability across users.

However, the learning formulation is particularly useful to accommodate selfish users. Non-cooperative

learning games have the advantage of very low overhead for information exchange in the network while the

cooperative learning game improves the overall network performance at the expense of increased overhead

required for information exchange. Note that the case of non-convex game is considered in [128].

In [129], the authors provide a comprehensive survey on the use of artificial intelligence for cogni-

tive radio networks. It reviews different implementations of cognitive radio which are designed based

on intelligent algorithms including artificial neural networks (ANNs), metaheuristic algorithms, hidden

Markov models (HMMs), rule-based systems, ontology-based systems (OBSs), and case-based systems

(CBSs). [129] also discusses the issues related to the selection of intelligent algorithms (e.g., responsive-

ness, complexity, security, robustness, and stability) for cognitive radio networks.

The work in [130] introduces an innovative method of designing cognitive radio by adopting the human

behavior model. The concepts of homo egaulis society, homo parochius society, and homo reciprocans

society in social science are discussed in the cognitive radio context with illustration. The work in [131]

uses the machine learning techniques in cognitive radio, in contrast to the traditional methods which rely

on the policy-based and hard-coded approaches and presents a concrete model of a generic cognitive radio

with learning engine.

The work in [132] considers the problem of aggregated interference management due to multiple

cognitive radios using ideas from artificial intelligence. The IEEE 802.22 standard is modeled as the

multiagent system where agents are the secondary base-stations supporting and controlling the data trans-

mission of secondary users. A real-time multiagent reinforcement learning algorithm (i.e., decentralized

Q-learning algorithm) is introduced to manage the aggregated interference. Both complete and partial

information cases are evaluated in which the optimal and suboptimal solutions are obtained, respectively.

The computational and memory requirements for implementations of these reinforcement algorithms are

evaluated.

The work in [133] adopts the concept of Bayesian nonparametric learning algorithm to optimize the

repeated spectrum auction in a cognitive radio network. The repeated spectrum auction is used which

considers the monitoring and entry costs of the secondary users to bid for the radio resource from the

Page 26 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

25

primary users. Since the knowledge of the secondary users about other users is limited due to the distributed

environment, a secondary user learns from experience using Bayesian nonparametric belief update scheme

and adapts the bidding strategies accordingly. Based on the belief, the secondary user decides to join the

auction or not. In addition, the bidding strategy is proved to be optimal.

In [135], the authors consider the multi-user distributed dynamic spectrum access problem in a spectrum

overlay scenario based on game-theoretic learning. In particular, a carrier-sense multiple access (CSMA)

protocol is considered and a game theoretic correlated equilibrium is achieved. In [136], the authors

consider cognitive radio networks for delay-critical applications using game theory and learning algorithm.

Previously, the cognitive radio is believed to be useful only for non-real-time applications (e.g., best-effort

applications). However, by using advanced techniques in optimization and artificial intelligence, this paper

proves the feasibility of deploying real-time applications (e.g., multimedia) in cognitive radio networks.

This paper presents centralized and decentralized spectrum market models based on the stochastic game

framework.

The work in [137] proposes the use of neural network from the field of artificial intelligence to implement

dynamic spectrum selection in cognitive radio networks. A neural network-based cognitive engine is

developed to learn from the environmental measurement on the different channels. Then, the algorithm

chooses the channel with the highest chance to yield the best performance based on the result from the

cognitive engine. The experimental measurement based on the testbed implementation is conducted to

evaluate the performance of the proposed cognitive radio design.

The work in [138] also adopts a neural network to predict the availability of the spectrum for cognitive

radio users. With a multilayer perceptron (MLP) neural network, where the training data is assumed to be

available, the accuracy of the spectrum prediction by the secondary user can be improved significantly.

Therefore, a secondary user can efficiently access the available spectrum without collision with the

transmission by the primary users. The authors in [139] investigate on distributed learning of channel

availability statistics and channel access in cognitive radio networks and present distributed channel access

policies based on the results on classical multi-armed bandit problem. For these policies, the bounds on the

regrets are obtained. The work in [140] presents a novel learning-based approach for dynamic spectrum

access by secondary users in a spectrum overlay scenario where the probability of collision with the

primary users needs to be bounded. This approach significantly outperforms the traditional listen-before-

talk approach.

Page 27 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

26

F. Multi-hop Transmission, Routing, Spectrum Mobility, and Cross-layer Optimization

In multi-hop cognitive radio networks [141], cognitive radio users communicate with each other in

a local area either in a centralized or an ad-hoc manner. Hence, routing selection plays an important

role in multi-hop cognitive networks. A local coordination based routing scheme for multi-hop cognitive

networks is proposed in [142]. In [143], joint routing, subband division and scheduling scheme for multi-

hop CRNs is considered. In [145], a cross-layer optimal scheduling algorithm is proposed for cooperative

multi-hop cognitive radio networks. In this network, secondary users help primary users to transmit in a

multi-hop fashion and the secondary users receive a spectrum opportunity for their transmission in return.

The analysis is developed to achieve optimal throughput for the primary users, where the upper bound is

derived.

To exploit the time-varying channel opportunities and provide QoS requirement to the users and save

transmission power, the authors in [148] consider dynamic rate allocation, routing and spectrum sharing

scheme for multi-hop cognitive networks.

In [143], the authors study the problem of multi-hop networking with cognitive radio nodes. For the

cognitive radio network, each node has a pool of frequency bands (typically of unequal size) which can be

used for data transmission. The difference in the bandwidth among the available frequency bands provokes

the need to divide these bands into sub-bands for optimal spectrum sharing. The authors characterize the

behaviors and constraints for a multi-hop cognitive radio network from multiple layers, including the

modeling of spectrum sharing and sub-band division, scheduling and interference constraints, and flow

routing. The objective is to minimize the required network-wide radio spectrum resource for a set of user

sessions. The problem can be formulated as a mixed-integer non-linear programming (MINLP). However,

this MINLP problem is NP-hard. Therefore, the authors develop a lower bound to estimate the objective

function by relaxing the integer variables and using a linearization technique. Simulation results show

that the solutions obtained by the proposed algorithm are very close to the lower bounds obtained via the

proposed relaxation. Therefore, the solution produced by the algorithm is nearly optimal.

The work in [122] studies the data link layer QoS performance measures for cognitive radio users in

an infrastructure-based dynamic spectrum access environment. The authors develop a queuing analytic

framework with a discrete-time Markov chain (DTMC) and derive important QoS measures. The model

considers primary users’ activity as a two state Markov chain, cognitive radio users’ channel activity as

a finite state Markov chain, bursty traffic arrival pattern at the cognitive radio user ends, and correlated

Page 28 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

27

channel fading. The authors provide a step-by-step procedure to derive the delay distribution, average

throughput, and packet loss rate for the cognitive radio users. The delay distribution can be derived by

considering an absorbing Markov chain. The proposed framework has important applications including

cross-layer analysis and design for improved QoS experience, and model-based admission control in

cognitive radio networks.

There have been only a few work on spectrum handoff and its effect on the performance of cognitive

radio networks. In [123], the authors outline the research issues and challenges related to spectrum

mobility in the context of spectrum management process for cognitive radio networks. The work in [124]

comprehensively models the effects of spectrum handoff and spectrum management methods on the call-

level QoS performance of secondary users. The work in [125] demonstrates spectrum handoff and spectrum

mobility features of cognitive radio (based on non-contiguous multi-carrier transmission) using software-

defined radio.

Robustness and security are important issues in the development of cognitive radio networks. In the

next section, we review the existing literature on security and robustness in cognitive radio networks.

V. ROBUSTNESS AND SECURITY OF COGNITIVE RADIO NETWORKS

A cognitive radio network could be vulnerable to different types of attacks, which could degrade

the performance considerably. [149] provides a comprehensive study and analysis on the security issues

in a cognitive radio network. The cognitive radio devices could wrongly learn from environment and

be taught by malicious users. Various attacks (e.g., sensory manipulation attacks against policy radios,

belief manipulation attacks against learning radios, and self-propagating behavior leading to cognitive

radio viruses) can use this approach. To address the problem, different methods are discussed. Firstly,

the robust sensory input can be performed by allowing the secondary users to run a cross-correlation

algorithm to determine the difference between noise and malicious signal. Secondly, all possible states of

the environment must be enumerated to obtain the complete knowledge so that the secondary user will

be aware of an attack. Thirdly, multiple secondary users can cooperate to form a group and exchange

information if there is an attack. Each of these methods may have some drawbacks. The cross-correlation

algorithm could be too complicated to implement in a typical secondary node. State enumeration requires

too much memory and too long execution time of a secondary user. Finally, forming a group and

exchanging information could incur an unacceptable overhead in a network. Alternatively, some defense

mechanisms for cognitive radios against the attack (e.g., using trust network) are also mentioned.

Page 29 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

28

In a cognitive radio system, it is important to distinguish between the signals from primary and secondary

users. Especially, in a hostile environment, the attacker may use a modified air interface to imitate the signal

of primary users, making secondary users unable to access the spectrum. [150] systematically shows that

the aforementioned attack can result in severe interference and significantly reduce spectrum utilization.

To solve the problem, [150] proposes a transmitter verification scheme, namely, the LocDef (localization-

based defense) scheme. This LocDef scheme is able to identify whether the signal is from a primary

user or not by using an estimated location of the transmitter (i.e., based on non-interactive localization

technique) and characteristics of the signal itself (i.e., received signal strength [RSS]). [150] considers the

log-loss signal propagation model, in which the expected RSS is given by µ = P + β0 + β1 ln d, where d

is the distance between actual primary transmitter and secondary receiver, P is the transmission power,

and β0 and β1 are constants for a particular environment. The value of RSS is then smoothed to extract

the important pattern, and the RSS peak can be used to identify the primary transmitter.

Adversaries

Emulated incumbent signal

Local sensing reports

Attacker

Falsified sensing result

Incumbent signal transmitter

(primary user) Data collector

Actual incumbent

signal

Final spectrum sensing result

Spectrum sensors

Fig. 2: Spectrum emulation attack.

In [151], the authors consider incumbent emulation and spectrum sensing data falsification attacks

(Fig. 2) which can degrade the performance of distributed spectrum sensing in cognitive radio networks.

In particular, an attacker (e.g., malicious secondary user) tries to gain priority over other secondary

users by transmitting signal emulating that of a primary user (i.e., incumbent emulation) or reporting false

spectrum sensing data (i.e., spectrum sensing data falsification). Defense methods against these attacks are

introduced. For the incumbent emulation attack, defense methods based on distance ratio test and distance

difference test are considered. The distance ratio test uses the RSS obtained from the pair of location

verifiers (e.g., dedicated spectrum sensor devices or secondary user devices) to verify the location of the

transmitter. However, the distance ratio test suffers from channel fading and shadowing, and may not yield

Page 30 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

29

the accurate distance to the transmitter. The distance difference test uses the relative phase difference of

the signal obtained at a pair of location verifiers to verify the transmitter. The distance difference test needs

synchronization among location verifiers to obtain the accurate result. For the spectrum sensing falsification

attack, the attacker tries to manipulate the spectrum sensing at the spectrum sensors or data collector (i.e.,

fusion center). Different methods related to the data fusion techniques are considered against the spectrum

sensing falsification attack, i.e., decision fusion, Bayesian detection [152], and Neyman-Pearson test. In

the decision fusion method, the sensing results from spectrum sensors are compared with the rule (e.g.,

threshold-based). In the Bayesian detection, the knowledge of a priori conditional probabilities of the

spectrum sensing is used to detect the attack. In the Neyman-Pearson test, the maximum acceptable

probability of false alarm or the maximum acceptable probability of misdetection is used to determine

the presence of an attack.

The work in [153] and [154] develop a defense strategy against the primary user emulation attack. The

passive anti-primary user emulation attack approach here is similar to the random frequency hopping used

for anti-jamming in traditional wireless networks. In this approach, the defender randomly switches the

channel to sense and access so that the defender can escape from the attack on a particular channel by the

attacker. In [153] (unlike that in [154]), the channel statistics (e.g., availability probability) are assumed to

be known. In [153], the attack and the defense mechanisms are modeled as a zero-sum game. The actions

of the attacker and the defender are the channel to jam and to perform channel sensing, respectively.

The payoff of the attacker is a constant reward if the attacker successfully attacks the defender, and

zero otherwise. Similarly, the payoff of the defender is the same constant reward if the defender can

successfully avoid the attack by the attacker, and zero otherwise. The Nash equilibrium is analyzed for

this game. The metric named anti-jamming efficiency (AJE) is introduced to measure the outcome of the

game, which is defined as follows:

AJE =Expected reward at the Nash equilibrium

Expected reward without jamming. (1)

The exact expression for anti-jamming efficiency is derived. The analysis in [153] also considers the

multiple stage case. The problem is simplified due to the partial observation of channel sensing and

imperfect monitoring. A POMDP is formulated and solved for the secondary user defender to optimally

sense and access the channel. The work in [154] extends the analysis of a primary user emulation attack

for the unknown channel statistics case. The secondary user as the defender must address the uncertainty

Page 31 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

30

in the channel statistics and time-varying policy of the attacker. The defender therefore has to learn and

adapt the channel selection policy accordingly.

In [155], the authors consider the problem of malicious secondary users deliberately sending false

reports in collaborative spectrum sensing. This is the spectrum sensing falsification attack. In a cognitive

radio system, the secondary users sense the channel independently. Then, the secondary users report to

the data collector where the OR rule is applied to determine the final spectrum sensing result. A method

based on the double-sided neighbor distance algorithm is proposed to detect attacks. A T -dimensional

space data, where T is the number of spectrum sensing periods, is used for detection. The distance (e.g.,

Euclidean, Hamming, and double-sided neighbor) is calculated from this historical sensing data, which

is used to identify the outlier secondary user far away from most secondary users in the history space.

Performance evaluation is performed to analyze the probability that the malicious user sends the sensing

result different from that of the secondary user.

In addition to the technical considerations, the design, optimization, and deployment of cognitive radio

networks will be affected by economic considerations. In particular, market mechanisms will be used

in designing efficient spectrum allocation and sharing methods. In the following section, we survey the

related work on different issues related to the economics of cognitive radio.

VI. ECONOMICS OF COGNITIVE RADIO NETWORKS

In a cognitive radio system, pricing is an important issue, which motivates the primary and secondary

users to share the available spectrum through a process often called spectrum trading [156]. Spectrum

trading is the mechanism for the entities in a cognitive radio system (e.g., primary and secondary users,

spectrum owner and users, service provider and subscriber) to exchange the radio resources. The exchange

could be performed through money or through different form of resources (i.e., bartering). In spectrum

trading, the primary users or service providers attempt to sell unused spectrum resources to secondary

users for monetary gains and the secondary users attempt to buy these resources to achieve their desired

communication goals. Two major approaches for spectrum trading are based on auction and open market.

A. Spectrum Auction

Auction is a traditional but efficient way to distribute commodity and services to the customers. The

theory of auction can be naturally applied to cognitive radio systems since the price of radio spectrum

cannot be determined precisely in advance [157]. In an auction, the primary users submit their asks and the

Page 32 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

31

secondary users submit their bids to sell and buy a radio resource, respectively. Different auction models

have been applied to cognitive radio (e.g., single-sided, double-sided, English, Dutch, Vickrey-Clarke-

Groves (VCG) and combinatorial auctions). In [158], an auction-based algorithm (second-price auction

mechanism) is developed, which helps the wireless users to compete for channels in a fair manner. In this

auction, the users bid for a channel given the channel state. The users can optimize their bids to maximize

the throughput under the budget constraint. Game theory is applied to analyze the bidding strategies of

users and it is shown that the Nash equilibrium exists which also leads to the unique allocation of channel

under certain channel state distribution. In [159], different auction models (i.e., single object pay-as-bid

ascending clock auction [ACA-S], traditional ascending clock auction [ACA-T], and alternative ascending

clock auction [ACA-A]) are proposed to sell the spectrum of a primary user to multiple secondary users.

A multimedia streaming application with delay requirement is considered for secondary users to bid for

the spectrum. All of the proposed auction algorithms (i.e., ACA-S, ACA-T, and ACA-A) are distributed

and all of them converge within a finite number of auction rounds.

[160] considers spectrum auction in which a regulator sells channels for primary and secondary

networks. The primary network on each channel has higher priority to transmit data than that of secondary

networks, while all secondary networks have the same priority. Two objectives are considered, i.e., to

maximize revenue of regulator (i.e., auctioneer) and to maximize social welfare (i.e., sum of the valuations

of all networks). Different algorithms are developed for these objectives. It is also shown that the optimal

channel access allocation algorithm of dependent bids (i.e., the network has some knowledge or belief about

transmission pattern of other networks) is an NP-complete problem. Then, for the case of independent bids,

the dynamic programming algorithm is developed to solve the channel access allocation problem. [161]

studies the truthfulness of secondary users to inform the valuation and arrival-departure time period in

the spectrum auction performed by the primary user. The spectrum has the expiry constraint, where the

spectrum is only available for a certain period of time. In this auction, the secondary users have an

incentive to misreport the valuation and arrival-departure time period to gain more spectrum allocation

from the primary users. Therefore, to avoid such misreporting, the channel allocation and pricing schemes

are introduced to minimize the impact of false information.

To adapt the bidding strategy competitively, the bidders (e.g., secondary users) must have knowledge of

other opponents. However, information about other bidders may not be available publicly and the bidders

must learn from past auction results and history. In a cognitive radio, this learning capability can be

implemented for auction [162], [163]. In [162], a learning algorithm is adopted to optimize the repeated

Page 33 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

32

spectrum auction in a cognitive radio network. A repeated spectrum auction is used which considers the

monitoring and entry costs of the secondary users to bid for the radio resource from the primary users.

Since in a distributed environment the knowledge of the secondary users about other users is limited, the

secondary users learn from experience by a Bayesian nonparametric belief update scheme and adapt the

bidding strategies accordingly. Based on the belief, the secondary user decides to join the auction or not.

In addition, the bidding strategy is proved to be optimal.

B. Price Competition in Open Market

Unlike an auction, which needs an auctioneer to control the trading, in an open market model, the

primary users and secondary users are free to sell and buy radio resources, respectively. Since there

is no control, the pricing strategy of primary users plays an important role here which determines the

revenue generated by the primary users. Also, pricing influences the decision of the secondary users to

buy the radio resource. A competitive pricing scheme based on a non-cooperative game among multiple

primary networks is proposed in [164]. A static game model and a dynamic game model are introduced

with and without global information, respectively. In the dynamic game model without complete global

information, the primary networks can adjust their pricing strategies to reach the Nash equilibrium. In

addition to non-cooperative pricing, the collusion among the primary users is analyzed, in which if the

collusion happens, the primary networks will adopt a cooperative pricing scheme instead of a competitive

one.

In [165], a joint power/channel allocation scheme is presented based on a distributed pricing approach.

The spectrum allocation is modeled as a non-cooperative game where the players are cognitive radio

users. A price-based iterative water-filling algorithm is introduced to reach the Nash equilibrium solution.

The game theory can model the behavior of cognitive radio users with self-interest, and therefore,

the game theoretic-models are very useful in modeling and optimization of dynamic spectrum sharing

networks [166]. With game-theoretic models, it is possible to design dynamic spectrum access methods

which can achieve flexibility, efficiency, and fairness. The work in [167] adopts a dynamic game model

to analyze competitive spectrum sharing. The importance of the work in [167] lies in the consideration

of dynamic behavior of strategy adaptation of primary users to sell the radio spectrum. In this case,

the convergence and the optimality of strategy adaptation are evaluated. The conditions to ensure the

stability of strategy adaptation are derived which will be useful for the primary users to avoid undesirable

fluctuations.

Page 34 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

33

Primary user 1

Spectrum F1 Spectrum F2 Spectrum FM

Primary user 2 Primary user M

A group ofsecondary user 1

Group AGroup 2

Fig. 3: Spectrum trading in cognitive radio with multiple spectrum sellers and buyers [168].

In the most general scenario, spectrum trading in a cognitive radio system may involve multiple spectrum

sellers and buyers. In [168], the authors develop a spectrum trading framework for this general spectrum

trading scenario (Fig. 3). An evolutionary game model is used for the spectrum selection of secondary

users when choosing the primary user (or service providers) to buy spectrum from. A hierarchical game

model is presented to obtain the equilibrium solution for the primary users in selecting the spectrum price

to maximize their profit given the performance degradation of the primary users due to sharing available

spectrum with the secondary users. The collusive behavior of users could be a significant threat to efficient

dynamic spectrum allocation in a distributed cognitive radio network. [169] presents a systematic approach

to avoid collusion in cognitive radio. In [169], the spectrum allocation with multiple selfish legacy spectrum

holders and unlicensed users is modeled as a multi-stage dynamic game, and a pricing-based distributed

collusion-resistant spectrum allocation approach is used to optimize overall spectrum efficiency. [170]

studies price-based spectrum access control in which the primary users sell spectrum to secondary users.

The secondary compete with each other through random access. Therefore, the concept of “price of

contention” is introduced. The decentralized pricing mechanisms are developed for the monopoly and

multiple primary user market. The solution in terms of a Nash equilibrium is considered. [171] considers

spectrum leasing problem between primary and secondary users. Noncooperative game is formulated to

determine the optimal parameters.

For the deployment of practical cognitive radio systems, development of simulation tools, testbeds, and

hardware prototypes are essential. The simulation tools provide the capability for the hardware, software,

and protocol developers, engineers, and researchers to test and evaluate the design and new concepts for

cognitive radios with low cost, small effort, and short time. In addition, the testbed and the hardware

platforms can be used to develop prototypes for investigation and experimentation of the actual cognitive

Page 35 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

34

radio systems. In the next section, we review the simulation tools, testbeds, and hardware prototypes

developed for cognitive radio systems.

VII. COGNITIVE RADIO SIMULATION TOOLS, TESTBEDS AND HARDWARE PROTOTYPES

A few work in the literature have introduced and reported simulation tools, testbeds, and hardware

prototypes for cognitive radio. [172] describes a simulation tool for performance evaluation and assessment

of a cognitive radio network with existence of multiple wireless systems (i.e., cellular networks and

opportunistic networks which are primary and secondary users, respectively). The simulation tool considers

the high-speed downlink packet access (HSDPA), hybrid ARQ (HARQ), and adaptive modulation and

coding (AMC) features for cognitive radios. [173] presents a cognitive radio simulation platform. The

customizable and extensible platform is developed based on C/C++ language. The major components in

the platform are the network topology, channel access of a primary user, mobility model, protocols (e.g.,

MAC and routing), traffic scheduling, and physical wireless channel. The platform has a set of performance

measures to be obtained, including throughput, spectrum efficiency, packet loss rate, jitter and delay.

[174] considers the issue of random number generation used for frequency-dependent spatial correlation

and position applicable for cognitive radio simulations. It is argued that the Kronecker approximation

cannot account for the dependencies. Therefore, a new method is introduced which uses nonassociative

left and right products of user-dependent frequency correlation matrices and frequency-dependent spatial

correlation matrices. While [174] concerns only random number generator, [173] provides a complete set

of simulation modules for cognitive radio.

Analog-digitalConverter (ADC)

Digitalprocessing

Digital-analogconverter (DAC)

Power amplifier

Isolator

Converter (ADC)

Fig. 4: Software-defined radio introduced by Mitola [176], [177].

[175] provides a comprehensive survey on the design methodologies for constructing flexible software-

defined radio receiver for cognitive radio based on the concept proposed by Mitola [176], [177] (Fig. 4).

The receiver should be able to receive data on a channel with any band from 800 MHz to 6 GHz and

any bandwidth. The design of the receiver is proposed to have a digital front-end which can be tuned

electronically. Then, the analog baseband samples the channel of interest at zero intermediate frequency

Page 36 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

35

(IF). The clock-programmable downsampling is applied with the embedded filtering. With this design, the

receiver can change the operating frequency band dynamically. The use of this design in CDMA requires

the system to operate in full-duplex mode (i.e., the transmitter and the receiver operate simultaneously),

and the ability of the receiver to operate on multiple channels concurrently.

[178] introduces a real-time cognitive radio testbed for physical and link layer experiments. This testbed

is based on the Berkeley Emulation Engine 2 (BEE2), which is a multi-FPGA emulation platform. Using

BEE2, the testbed can connect to 18 radio front-ends, and these front-ends can be set to be the primary

and secondary users in various test scenarios. In addition, by using FPGA, the testbed can simultaneously

operate on multiple radios, which is suitable for complex experiments and performance studies involving

high-speed low latency links. [179] discusses the design and implementation of an UWB cognitive radio for

high data rate communications with very low transmission power. The design of ultra-low power baseband,

impulse-UWB transceiver front-end, and digital backend implementation is introduced. In addition, [179]

reports a testbed implementation based on the BEE2. In [180], the design of a software-defined radio from

a low-power ADC perspective is introduced. The design is based on the windowed integration sampler

and clock-programmable discrete-time analog filters. By using a low-noise amplifier (LNA) and a wide

tuning-range synthesizer, the wideband RF front-end can operate on any frequency between 800 MHz

and 6 GHz. The wideband LNA can achieve a good performance (e.g., 18-20 dB of maximum gain). The

performance evaluation is performed on the GSM and the IEEE 802.11g standards.

There are a number of wireless testbeds with cognitive networking capabilities which have been

developed in the academia and they differ in terms of the scale of the network, the degrees of experimental

control and isolation they offer, and the indoor and/or outdoor nature of the deployment and measurement

environment [3]. They are based on the off-the-shelf radio technologies such as Wi-Fi and WiMAX

technologies. A variety of software radio platforms for base stations, client nodes, spectrum sensors,

and different RF sources or users are used in these testbeds. These platforms are generally capable of

implementing different modulation algorithms, MAC protocols, and adaptation methods. Examples of

the testbeds include the CMU emulator [181], the ORBIT radio grid testbed [182], the “DieselNet”

testbed [183], and the “CitySense” network [184]. The CMU wireless network emulator testbed has 15

nodes (with 802.11b or Bluetooth interfaces) and it supports the entire 2.4 GHz ISM band. The nodes

communicate through the emulator, which is an FGGA-based central DSP engine, and this emulator

models the effects of wireless signal propagation. The ORBIT radio grid testbed, which contains 400

programmable radio nodes in an area of 5000 sq-ft, is also based on an emulator, and the testbed can

Page 37 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

36

be accessed through an Internet portal. It supports end-to-end wired and wireless experiments. Also, the

testbed includes an outdoor field trial system. Examples of cognitive radio platforms include WiNC2R

cognitive radio board [185], the Kansas University Agile Radio (KUAR) [186], the Defense Advanced

Research Projects Agency (DARPA) Wireless Network after Next (WNaN) cognitive radio [187], the

Universal Software Radio Peripheral (USRP) and GNU radio software architecture [188], and the RICE

WARP board [189]. In general, these platforms consist of RF transceiver modules, digital processing

boards, and networking interfaces/modules, and they are supported by software architectures such as

GNU.

VIII. FUTURE TRENDS IN COGNITIVE RADIO NETWORKING RESEARCH AND OPEN ISSUES

In this section, we summarize future trends in cognitive radio networking research which include

the cooperative and cognitive radio communications, cognitive radios for the emerging LTE-Advanced

networks, and cognitive radio applications. Also, the major open research issues are outlined.

A. Cooperative and Cognitive Communications

Relay-based communication is a powerful approach to combat channel fading, extend transmission

coverage, and improve system capacity [190]. This approach can be also used for cognitive radio net-

works to improve the network capacity, reduce interference to primary users, and achieve fairness in

data transmission among the cognitive radio users [191]–[196]. However, due to the different spectrum

availability at the cognitive radio nodes (i.e., transmitter, receiver, and relay) in a cognitive relay network,

the traditional relay-based communication approaches cannot be directly applied.

The relay-assisted cognitive transmission can be divided into two main scenarios, i.e., cooperative

transmission between secondary users [197], [198] and cooperative transmission between primary users and

secondary users [191], [199]. In the first scenario, the secondary users act as a relay to assist only secondary

user’s transmissions. In the second scenario, each secondary relay cooperates by relaying primary user’s

signal to its primary receiver, and as an incentive, each secondary relay gets an opportunity to transmit

in the same spectrum which is used by the primary user. One of the most difficult challenges in relay-

based cognitive systems is the design of MAC protocols which are able to utilize the best relay and

the spectrum opportunities for the optimal performance. A centralized relay assisted MAC protocol is

presented in [191], [194] for relay-assisted cognitive networks.

Page 38 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

37

For resource allocation in relay-based cognitive radio networks, in addition to power and rate allocation,

the channel and relay allocation are also important since the secondary users as well as the relays contend

for channel access. Thus, the allocation of both relay and channel resource is coupled together and must

be addressed jointly [197], [198], [200]–[205]. The selective fusion spectrum sensing and best relay data

transmission (SFSS-BRDT) scheme and the fixed fusion spectrum sensing and best relay data transmission

(FFSS-BRDT) scheme are used to address this relay selection problem [206]. In [207]–[209], centralized

solutions are considered in which the resource allocation at the transmitter (cognitive radio base station)

and at the relay stations is computed in a centralized manner based on the global channel state knowledge.

Due to the signaling overhead and computational complexity, decentralized resource allocation is more

desirable for cognitive relay networks [210]. In practical implementation, due to imperfect channel sensing,

the relay node may not always be able to acquire available spectrum for relay transmission. Further, due

to the imperfect channel state information (CSI), the selected relay may not always be the best relay. The

problem of optimal resource allocation under imperfect spectrum sensing and imperfect CSI in cognitive

relay networks has been addressed in the recent literature [199], [210]–[212].

B. Cognitive Radio for LTE/LTE-Advanced Networks

Dynamic resource management is the one of the most significant features in cognitive and software-

defined radio systems. Future LTE/LTE-Advanced networks will have to address more complex resource

management problems involving both contiguous and non-contiguous carrier aggregation, dynamic co-

channel and adjacent channel interference mitigation, advanced multiple-input multiple-output (MIMO)

techniques, multiple radio access technology (RAT) functionalities, coordinated multiple point (CoMP)

transmission and reception, and relaying for efficient use of radio spectrum [213]. Cognitive radio concepts

will be useful to develop enabling techniques to implement these features in the LTE/LTE-Advanced net-

works. LTE-Advanced networks will also use the concept of femtocells (also referred to as heterogeneous

networks) to provide higher quality of service to the end users by improving the indoor coverage. However,

deployment of femtocells will result in a complex interference scenario when compared to that due to

traditional frequency planning. Interference mitigation in such a scenario will require frequency agility,

dynamic channel selection, and complex power control. The cognitive radio technologies can be used in

the heterogeneous networks to enable the femtocells to perform the required adaptation.

Page 39 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

38

C. Cognitive Radio Applications

1) General applications: A comprehensive survey of the emerging cognitive radio applications is

presented in [214]. This paper identifies some killer applications of cognitive radio technology and

discusses about the opportunities. The major applications include smart grid communications, public

safety communications, and broadband cellular to medical applications. This paper discusses how to

adopt cognitive radio in these applications. The benefits and the challenges are outlined and the possible

solutions are discussed. Last but not the least, this paper reviews some standardization activities related

to adopting cognitive radio in various applications.

The potential applications of cognitive radio to the public safety communications are discussed in [215].

The public-safety community will adopt dynamic spectrum access technique to improve the efficiency,

capacity, robustness, and flexibility of the wireless communications systems. This paper raises an important

issue regarding such an adoption which is the spectrum management reform to create spectrum pools.

With these spectrum pools, cognitive radio devices can utilize the spectrum efficiently under the control of

adaptive prioritization policies. This approach overcomes many limitations in the legacy communications

systems. This paper also discusses about the economic, policy, and business issues of such spectrum

pooling.

The work in [216] discusses the use of cognitive radio in the IEEE 802.16-based WiMAX networks.

Although the original IEEE 802.16 standard does not adopt any cognitive radio into account, this paper

points out that significant benefit can be achieved by increasing frequency reuse through power control and

cognitive channel assignment. Then, the paper introduces an algorithm for optimal channel assignment.

Simulations are performed for an IEEE 802.16 system to verify the benefits of the proposed algorithm.

One of the very promising applications of cognitive radio will be its use in the evolving femtocell

networks to improve the indoor coverage as well as the capacity of traditional cellular wireless networks.

The work in [217] introduces an interesting idea of ultra-broadband femtocells based on the cognitive radio

concept. Specifically, the opportunistic spectrum reuse method is presented which allows the femtocells

to access the radio spectrum licensed to different operators with different services. The architecture

and the key technology enablers are discussed to achieve the optimal performances of heterogeneous

networks with multiple macrocells and femtocells. The work in [218] proposes a femtocell-based cognitive

radio architecture to enable multitiered opportunistic access in next-generation heterogeneous networks

(HetNets). The architecture unifies the concept of conventional femtocells with infrastructure-based overlay

Page 40 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

39

cognitive radio networks.

The principles of cognitive radio can be also used in multihop wireless mesh networks. Instead of a

fixed channel allocation among the mesh nodes, a dynamic channel allocation can be used to improve

the spectrum utilization. With this spirit, in [219], a cluster-based dynamic channel allocation framework

is proposed for cognitive wireless mesh networks taking the issues of interference and coexistence with

primary users into account. In [220], the authors present applications of cognitive radio in the context

of green communications. Two approaches are discussed, i.e., applying cognitive radio to achieve energy

efficiency and developing energy efficiency for cognitive radio systems. The paper introduces different

techniques for achieving energy efficiency in wireless systems and how to apply those techniques in

cognitive radio.

2) Cognitive radio networking in the TV white space: Cognitive radio networks can be developed to

exploit the TV white space (TVWS) for broadband wireless access, as the analog TV bands, which is

largely underutilized, has been released for data communications. Many studies in the literature consider

cognitive radio on the TV white space (e.g., [221]–[224]). In the TVWS, the TV broadcasting stations

and low-power wireless microphones serve as incumbents and the secondary systems such as the IEEE

802.22-based wireless regional area networks, the Wi-Fi hotspots and home networks can coexist with

the incumbent systems.

[221] provides a comprehensive survey on accessing TVWS. Various issues are identified, including

high-precision spectrum sensing [225], agile modulation and spectrum pooling techniques, and system-

level design. Also, the potential of using TVWS for a wireless home network and cognitive femtocell is

discussed. The coexistence challenges for heterogeneous cognitive networks in the TVWS were discussed

in [222].

In [223], the problem of coexistence among wireless networks in TV white space is studied. The

framework to exchange information related to coexistence is also introduced. This framework supports

both centralized and distributed information exchanges. The framework uses a multiradio cluster-head

equipment (CHE) to collect information about TV white space and coexisting wireless networks. In the

centralized approach, there is a central coexistence database available for the CHE to obtain information

about wireless environment and to make the decision on spectrum access. On the other hand, in the

distributed approach without the central database, a broadcast channel is used to exchange the coexistence

information among CHEs. The simulation results show that with the proposed framework, both centralized

and distributed approaches can discover larger spectral availability than that without information exchange.

Page 41 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

40

In [226], a wireless hardware prototype to operate on TV white space is introduced. The prototype has

a sensor and a geolocation device to collect information about the wireless environment in TV bands. This

information is stored in the TV band database. A cognitive management entity is introduced to process

the channel information and to control the transmission and reception of the transmitter and receiver,

respectively. The prototype is developed based on the TV white space test trial regulation in Singapore

which requires the wireless networks to operate on the 630-742 MHz band. The prototype is capable to

automatically search for and access unused TV bands.

[227] presents the feasibility study of using TV white space for wireless home networks. Also, the

performance is compared with that of 2.4 GHz and 5 GHz ISM bands which are currently widely used.

The simulation results show that the performance of using TV white space is better than that of 5 GHz

and 2.4 GHz bands for low to medium traffic load (e.g., several Mbps). Also, low power level (i.e., below

3 dBm per channel) can be used. However, at high traffic load, it is suggested that the TV bands should

be used with the interference management, and are suitable as the complementary channels of the 5 GHz

and 2.4 GHz bands.

3) Cognitive radio for smart grid communications: The concept of smart grid has been introduced

to improve the efficiency and responsiveness of the traditional power grid. The smart grid will use data

communications technologies to transfer different information related to the generation, transmission, and

distribution operations of the power grid to achieve the optimal operation of the power system. Wireless

technologies will be used for the smart grid communications [228]. Cognitive radio will be a potential

technology to improve the spectrum utilization and transmission efficiency of wireless communications

used in the smart grid [229]–[231].

In [229], the feasibility of using cognitive radio in the smart grid applications is extensively investigated.

The system architecture, algorithms, and hardware testbed are developed and analyzed for a microgrid

environment to support power and information flows. In addition, the mechanisms to ensure secure control

and information exchange in the power system are also considered. The independent component analysis

(ICA) together with the robust principal component analysis (PCA) techniques are used to improve the

robustness of the smart meter data transmission under strong interference. In [232], a hierarchical cognitive

radio communications infrastructure is proposed as a solution to manage a large number of transceivers

in the smart grid. In this infrastructure, cognitive radio will be potentially deployed for the home area

networks (HANs), the neighborhood area networks (NANs), and the wide area networks (WANs). Dynamic

spectrum access and sharing is also proposed to optimize the performance in each network to achieve the

Page 42 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

41

globally optimal spectrum access strategy. In [233], a cognitive radio network is developed for HANs.

The spectrum sensing and channel switching techniques are introduced for the smart meter to transmit

the meter data on either the licensed or the unlicensed channels. The objective is to improve reliability

and reduce delay in data transmission. The sensing time is optimized to reduce packet loss rate and delay

while the transmission of primary user is not interfered.

Although the above work addresses different issues of using cognitive radio in the smart grid, there are

many research opportunities. For example, a QoS framework for the cognitive radio used in smart grid

must be designed and optimized to meet the delay and reliability requirements of the different smart grid

applications. The benefits of cognitive radio in power generation, power transmission, and consumption

must be analyzed. Also, the integration of the cognitive radio into existing wireless communications

infrastructure of the smart grid needs to be investigated.

4) Cognitive radio for machine-to-machine (M2M) communications: Machine-to-machine (M2M) com-

munications allows wireless nodes to communicate without or minimal interaction with human. In M2M

communications, the data to be transmitted is smaller in size but more in frequency than those of the

human-to-human (H2H) communications. Cognitive radio is a potential solution for M2M communications

by offloading the data transmission from licensed channel to unlicensed channel. However, there are

many issues that need to be addressed. M2M communications will be application centric, in which the

protocol design and optimization must take the requirements of the specific applications into account. For

example, [234] considers M2M communications for the smart grid applications focusing on the energy

consumption of data transmission through cognitive radio networks (i.e., TV band devices [TVBDs]). The

dynamic spectrum access is optimized using a multi-objective genetic algorithm to find the optimal tradeoff

between power efficiency and spectrum efficiency. Also, the spectrum allocation for M2M communica-

tions must be determined. [235] studies the cognitive radio network tomography and spectrum map for

cooperative spectrum sensing. A two-step cooperative learning algorithm is proposed to analyze the spatial

correlation of the results from sensing nodes and the spectrum map is efficiently constructed. The proposed

algorithm is suitable for large-scale, dynamic, and heterogeneous M2M networks. M2M communications

can be integrated into the existing wireless systems (e.g., WiMAX and LTE [236] networks). However,

optimal system architectures must be developed for integrating M2M communications and cognitive radio

into such wireless systems. With the dense deployment of D2D communications in the limited cellular

radio spectrum, cognitive radio-based sophisticated interference control and avoidance methods will be

essential for these systems.

Page 43 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

42

5) Cognitive radio and mobile cloud computing: Mobile cloud computing combines wireless technolo-

gies and cloud computing to improve the performance and reduce cost of mobile services to the users [237].

In mobile cloud computing, the mobile application is divided into two parts, i.e., local computing modules

to run on a mobile device and remote computing modules to run on a server in the cloud to reduce the

energy consumption of a mobile device and to improve the performance of a mobile application. These

two parts need to communicate with each other through a wireless connection. Cognitive radio can be

used to support mobile cloud computing, i.e., to transfer data among local and remote modules of mobile

applications. The major challenge in using cognitive radio for mobile cloud computing is at the problem of

guaranteeing QoS to the users and efficient radio resource allocation [238]. The radio resource allocation

must be performed jointly with the computing resource (e.g., CPU of a server) allocation since mobile

applications can be executed in different modes, each of which requires different amount of bandwidth

and computing resource. Energy efficiency will be also an important factor which has to be optimized for

cognitive radios used in a mobile cloud computing scenario.

Cloud computing can be used to support cognitive ratio networks. Complex signal processing for

wireless transmission can be offloaded from mobile devices and base stations to the servers in the cloud.

This concept has been used in a few work in the literature. In [239], cloud computing is used as the

central communication and optimization infrastructure for a cognitive radio network. This network is

used for advanced metering infrastructure (AMI) to support the smart grid applications. In this case, the

data processing of AMI is performed by servers in the cloud. In [240], cooperative spectrum sensing

and radio resource scheduling for cognitive radio communications in the TV white space is implemented

and executed in a cloud computing environment. In particular, the spectrum sensing reports are analyzed

in the cognitive radio cloud (CRC) to identify the spectrum opportunities. The analysis is done through

the sparse Bayesian learning algorithm using Microsoft’s Windows Azure Cloud platform, which is one

of the public cloud computing services. In [241], the authors present a prototype system of a cognitive

radio architecture implemented using cloud computing to store and process spectrum sensing data. Cloud

computing is used for data sharing among network devices in the network centric operation.

D. Open Issues

The emerging cognitive radio technology is envisioned to become a key element in the future generation

“smart radio” technology which will dynamically optimize usage of limited radio resources and mitigate

the “radio traffic jam” problem in future wireless networks. This technology will also provide robustness

Page 44 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

43

and efficiency for wireless and mobile communications services. It is anticipated that cognitive radio will

evolve as a general-purpose programmable radio, which can be implemented through SDR, and it will serve

as a universal platform for flexible and adaptive wireless system developments [3]. Although significant

amount of research efforts have been spent on different aspects of cognitive radio communications and

networking, there are still many technical challenges and interesting research opportunities. Also, there

are significant challenges in the spectrum policy domain and pricing for dynamic spectrum access. In

addition, rigorous testing and evaluation of cognitive radio technology in real-world environments is a

significant issue which must be addressed to study the emergent behavior of cognitive radio and for

large-scale deployment of this technology. A few of the open technical challenges are outlined below.

1) Practical methods to estimate channel parameters and spectrum sensing for cognitive radios: A

cognitive radio system is characterized by a time-dependent structure of the primary user traffic pattern and

imperfect spectrum sensing. It has been investigated by a large body of cognitive radio literature, especially

from the standpoint of the decision-theoretic framework such as a partially observable Markov decision

process (POMDP), quickest detection, sequential detection, a restless bandit problem, and sequence

detection. In most of the existing work in the literature, secondary users’ knowledge of the channel

parameters is simply assumed without giving a specific method for obtaining the information. However,

in a real scenario, it may be the case that the primary user does not explicitly help the secondary user

in obtaining the knowledge of the channel parameters. Therefore, developing a method for the secondary

users to estimate the channel parameters without any help from the primary user is crucial for the design of

a practical cognitive radio system. Although a significant amount of research have been done on spectrum

sensing based on transmitter detection, there are only a limited number of techniques available for receiver

detection.

2) Practical spectrum sharing protocols for cognitive radios: Practical methods for spectrum sharing

among cognitive radios need to be developed considering the limited (or no) availability of control

channels for coordination of spectrum access in a multi-user, multi-channel, and multi-radio scenario.

Both horizontal and vertical spectrum sharing scenarios need to be considered. In this context, distributed

adaptation and resource management techniques, which depend on local information only and incur low

control overhead and implementation complexity, are desirable. For this, different machine learning, pattern

recognition, and artificial intelligence tools will be useful. In order to make the developed resource

allocation (i.e., power and rate allocation) methods amenable to real-world environments, they have

to be made robust against the various limitations encountered in practice such as the unavailability

Page 45 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

44

of instantaneous channel gain information (due to the time-varying nature of wireless channel) and

unavailability (or limited availability) of a control channel.

Also, practical protocol designs to support spectrum sharing using advanced communications techniques

such as network coding, multiple-input multiple-output (MIMO) communication, relay-based cooperative

communication in cognitive radio networks are interesting research problems.

Since the cognitive radio systems could be vulnerable due to the malicious eavesdroppers who try to

overhear other users’ messages. Therefore, spectrum access and sharing methods are required which can

maximize the secrecy rate (i.e., rate of perfectly secure transmission of information) of cognitive radio

users against the malicious behavior of eavesdroppers.

3) Protocol architectures and cross-layer adaptation: Design and evaluation of higher layer protocols

such as the transport layer protocols for cognitive radio networks in a wired-cum-wireless scenario have

not been adequately addressed in the literature. Design and analysis of transport and routing layer protocols

for a multi-hop or ad hoc cognitive radio network [242] also need extensive investigation. Cross-layer

adaptation and optimization will be a key component in the design of cognitive radio protocols for adaptive,

flexible, and stable dynamic spectrum access networks (Fig. 1). This optimization will need to consider

spectrum availability and dynamic channel and network selection in addition to the network parameters

such as QoS requirements, mobility, traffic load, node density, channel parameters as considered for

traditional cross-layer optimization in wireless networks. Also, a management and control framework will

be required for exchange of inter-layer information to support cognition and achieve cross-layer adaptation.

Design and evaluation of such frameworks is an open problem.

4) Spectrum policy research on cognitive radio: Spectrum policy research is required to define the

etiquettes for the cognitive radios, analyze their impacts on the technology and business strategies, analyze

their economic and social benefits, develop mechanisms to enforce the etiquettes derived from the spectrum

policy, and define the role of the different stakeholders [3]. Development of policy and legal frameworks

for cognitive radios for the different spectrum sharing models (or market models) is a very significant

research direction. Techniques from microeconomic and game theory models and incentive mechanisms

are useful in this context. Although there have been some work in this direction, more work are required to

analyze large-scale spectrum markets in terms of efficiency and convergence and understand their emergent

behavior as well as practical implementation of the derived methods and policies in the cognitive radio

protocols. Enforcement of the spectrum policies for the cognitive radios is another big challenge.

Page 46 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

45

5) Development of cognitive radio testbed, experimentation, and application-centric performance eval-

uation: Since cognitive radio networks are very complex systems with many degrees of freedom, it is very

critical to evaluate their system-wide performance in real-world environments and study their emergent

behavior to assess the viability of their practical implementation. Therefore, development of prototypes and

testbeds and their management are essential so that experimentations can be done in realistic environments

to complement the analytical models and simulations. The adaptive protocols and the policy requirements

add challenges for the testbed development and experimentation for cognitive radio networks. Again,

application-centric evaluations (e.g., for public safety applications, broadband Internet access in rural

areas, smart grid communications, eHealth applications, vehicular networking applications) of cognitive

radio systems are interesting avenues of future research on cognitive radio.

6) Cognitive small cell networks: Deployment of small cells has been identified as a potential approach

to increase the cellular wireless network capacity. By connecting to a small cell base station (e.g., a pico

or femto base station for picocell and femtocell, respectively), users can locally transmit and receive data

with much better channel quality. However, the small cells will be underlaid in a macrocell. If the small

cells and the macrocell use the same spectrum, interference will occur, deteriorating the performances of

users in both cell tiers. Cognitive radio techniques can be adopted to mitigate this problem. In this case,

the small cells can reuse the spectrum allocated to the macrocell. By opportunistic access, the spectrum

utilization can be improved while the network capacity is enhanced due to macrocell offloading [243].

For efficient opportunistic spectrum access, spectrum sensing mechanism to identify spectrum availability

needs to be carefully and optimally designed [244].

7) Cognitive Device-to-Device (D2D) communications: D2D communications has been introduced to

offload data traffic from a base station by allowing D2D devices to communicate with each other directly.

Similar to cognitive small cell networks, D2D communications can reuse the spectrum allocated to the

cellular users (i.e., users connected with the base station). Cognitive radio techniques can be applied to

D2D devices (i.e., secondary users) to opportunistically access the spectrum. Again, the spectrum sensing

is important to ensure that QoS requirements of the cellular users will not be violated while D2D devices

can achieve satisfactory performance.

8) Energy-efficient cognitive radio: Energy is an important resource for mobile devices [245]. There-

fore, techniques in cognitive radios (e.g., spectrum sensing [246], dynamic spectrum access [247] [248] and

sharing [249]) can be also optimized to maximize energy efficiency for data transmission. One approach for

spectrum sensing is to use an adaptive algorithm to perform sophisticated signal detection and estimation,

Page 47 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

46

which consumes a lot of energy, only when needed (e.g., a primary user is likely to present). For the

spectrum access, a secondary user must optimize the data transmission given the available energy [250].

Different forms of learning and reasoning techniques can be exploited for this. In the context of both single-

tier and multi-tier cellular wireless networks, cognitive radio techniques (e.g., for traffic load sensing and

estimation) can be used to opportunistically switch on/off some of the base stations while satisfying the

users’ communications requirements. Also, cognitive radio users can use energy harvesting techniques to

avoid relying on a fixed source of energy [251].

IX. COGNITIVE RADIO STANDARDIZATION

The IEEE 802.22 and SCC 41 are considered as the primary cognitive radio standards today. They are

also the completed standards of interest for cognitive radio [252]. However, there are also many other

standards being developed. IEEE created the 802.22 Working Group (WG) for WRANs in November,

2004. This WG has been assigned to develop an air interface (i.e., PHY and MAC) based on CRs for

unlicensed operation in the TV broadcast bands. The focus of 802.22 is on rural broadband wireless access

and its coverage range is considerably larger than that of the IEEE 802.16 [253]. A brief review of the

different standardization efforts is provided below.

A. IEEE 802.22

An overview of the 802.22 architecture (e.g., topology, entities, and connections), its requirements (e.g.,

service capacity, service coverage, physical and MAC layer details), applications, and coexistence issues

(e.g., antenna, TV and wireless microphone sensing and protection) are discussed in [254], [255]. In North

America, the frequency band of operation of the IEEE 802.22 networks is 54-862 MHz. Also, the standard

shall accommodate various international TV channel bandwidths of 6, 7, and 8 MHz. The IEEE 802.22

systems have a fixed point to multipoint air interface. The base station controls the consumer premise

equipments (CPEs). The base station also coordinates “distributed sensing” which is needed for incumbent

(original user) protection. The base station directs the CPEs to measure different TV channels and provide

feedback. For an average spectral efficiency of 3 bits/sec/Hz in the IEEE 802.22 systems, considering 12

users, the minimum peak throughput rate is 1.5 Mbps per CPE in the downstream direction and 384 kbps

in the upstream direction. Another striking feature of the IEEE 802.22 WRAN as compared to the existing

IEEE 802 standards is the base station coverage range, which can be up to 100 km if transmission power

is not an issue (current specified coverage range is 33 km at 4 watts CPE EIRP).

Page 48 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

47

B. IEEE 1900 - SCC41- DYSPAN

The IEEE Standard Coordinating Committee 41 (SCC41), formerly known as the IEEE 1900 task

force [256], was created by the IEEE to work in the area of dynamic spectrum access networks for

cognitive radio standardization. The IEEE SCC41 is divided into four Working Groups (WGs) termed

as 1900.x, “x” being the WG. In 2010, the SCC41 was approached by IEEE Communications Society

Standards Board (CSSB), and was renamed as IEEE DYSPAN-SC.

IEEE P1900.1 (Terminology and Concepts for Next Generation Radio Systems and Spectrum Manage-

ment) intends to create a glossary of important cognitive radio terms and concepts related to spectrum

management, policy-defined radio, adaptive radio, software-defined radio, and related technology, and also

compare different technologies and their capabilities [257].

IEEE P1900.2 (Recommended Practice for Interference and Coexistence Analysis): The 1900.2 WG rec-

ommends interference analysis criteria and develops a structure for measuring/analyzing the interference.

This standard provides a systematic way of analyzing interference and coexistence.

IEEE P1900.3 (Dependability and Evaluation of Regulatory Compliance for Radio Systems with Dy-

namic Spectrum Access): On the software side, the 1900.3 WG is developing test methods for evaluating

SDR devices. The main goal is to verify the coexistence and compliance of the software modules for

cognitive radio devices before certifying final devices.

IEEE P1900.4 (Architectural Building Blocks Enabling Network-Device Distributed Decision Making

for Optimized Radio Resource Usage in Heterogeneous Wireless Access Networks): This standard is for

radio systems with multiple radio access technologies (RATs) [258], [259]. The end terminal users are

considered to be supporting multiple RATs and with some cognitive radio capabilities such as flexible

operations in different frequency bands. The IEEE 1900.4 defines reconfiguration management entities

(RMEs), which help in decision making at terminal and network sides.

IEEE P1900.5: The SCC41 established this working group to address the policy language issues.

IEEE P1900.6: The SCC41 established this working group to address RF sensing functions that can

be managed in a terminal reconfiguration manager (TRM) defined in the IEEE P1900.4. Published in

April 2011, this standard provides definition of structure and interfaces for exchanging sensing related

information between sensor and users of sensing information (cognitive engines).

IEEE P1900.7 (Radio Interface for White Space Dynamic Spectrum Access Radio Systems Supporting

Fixed and Mobile Operation):

The standard would focus on the development of cost-effective white space dynamic spectrum access

Page 49 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

48

radio systems capable of operation in white space frequency bands on a non-interfering basis to incumbent

users in these bands.

The IEEE SCC41 has also started cooperating with the FCC, Federal office of Communication (OF-

COM) UK, SDR Forum and OMG forum, etc.

C. IEEE 802 Projects Related to Cognitive Technologies

IEEE 802.16.2: Completed in November 2001, it is one of the first coexistence standards which provides

guidelines for minimizing interference in fixed broadband wireless access (BWA) systems.

IEEE 802.15.2: Completed in June 2003, this standard outlines recommended practices for coexistence

of IEEE 802.15 wireless personal are networks (WPANs) with other wireless devices operating in unli-

censed bands. It also provides a technique for the IEEE 802.11 devices to coexist with the IEEE 802.15

devices.

IEEE 802.11h: Completed in September 2003, this standard provides mechanisms for dynamic frequency

selection (DFS) and transmit power control (TPC). The DFS feature developed for the IEEE 802.11 allows

relocation of interfering IEEE 802.11 basic service set (BSS) to another frequency. This standard has been

superseded by IEEE 802.11-2007.

IEEE 802.11y: This amendment to the IEEE 802.11 standard will allow 802.11 to operate at 3650-3700

MHz bands in the U.S. It will allow shared 802.11 operations with other users. It also includes transmit

power control (extending 802.11h) and DFS (extending 802.11h).

D. Wireless Innovation Forum (SDR Forum)

In 2008, the SDR forum published a report [260] on “Cognitive Radio Definitions and Nomenclature”.

This report highlights the components involved and defines many working technologies involved with

cognitive radio. The Cognitive Radio Working Group (CRWG) under SDR forum is also working on a

report entitled “Quantifying the Benefits of Cognitive Radio” developed for the benefits of using cognitive

radio technologies in upcoming wireless systems. In 2009, the SDR forum also started a test and measure-

ment project to develop a set of use cases, testing requirements, and methods for SSDR/CR technologies.

The project addresses critical TVWS functions, e.g., spectrum sensing, interference avoidance, database

performance, and adherence to policy.

Page 50 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

49

E. ETSI Standardization

In October 2009, ETSIs Technical Committee (TC) on Reconfigurable Radio Systems (RRS) released

a series of ETSI Technical Reports (TR 102 838) [261] that examine the standardization needs and

opportunities related to reconfigurable radio systems including SDR and CR. The reports summarize the

feasibility studies carried out by the committee and present the recommended topics for standardization.

Following the completion of the feasibility studies, standardization of RRS is now getting underway with

four different working groups studying different aspects of reconfigurable radio systems as follows:

System Aspect Studies (WG1): The basic role of this group is to form a technical framework that can

map other topics being studied by other groups.

Reconfigurable Radio Equipment Studies (WG2): This group finalized two study reports identifying

potential SDR standardization Mobile Devices (MD) and Reconfigurable Base Stations (RBS).

Cognitive Network Management Studies (WG3): This study is divided into two parts, i.e., Functional

Architecture (FA) and Cognitive Pilot Channel (CPC). The scope of FA is to study and propose a generic

architecture for the management and control of reconfigurable radio systems, namely, the Functional

Architecture (FA). The FA is designed to improve the utilization of spectrum and radio resources.

Public Safety Studies (WG4): This group studies the application of RSS to the public safety domain

in two parts, i.e., user requirements and operational requirements.

F. International Telecommunication Union (ITU) Standardization

The ITU-Radio communication sector (ITU-R) study Group 8 (Mobile, Radio determination, Amateur

and Related Satellite Services) deals with standardization of cognitive radio networks. ITU-R study Group

8 has published two reports on SDR [262], [263]. The reports focus on the application of SDR in

International Mobile Telecommunications-2000 (IMT-2000) technology. IMT-2000 systems are the third

generation mobile systems, which provide access to a wide range of telecommunication services, supported

by the fixed telecommunication networks (e.g., PSTN/ISDN/IP), and to other services, which are specific

to mobile users. The SDR technology is employed in the base station and the controllers of a mobile

radio access network (RAN) to increase the flexibility of radio access networks.

X. CONCLUSION

Cognitive radio represents a new paradigm for designing intelligent wireless networks to mitigate

the spectrum scarcity problem and provide significant gain in spectrum efficiency. We have provided

Page 51 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

50

a comprehensive survey of the research activities in cognitive radio. The major issues in the design of

cognitive radio communication networks have been discussed (e.g., spectrum sensing, dynamic spectrum

access, applications, and standardization) and the related work in the literature have been reviewed. First,

the historical note of the cognitive radio has been given to provide a motivation of the dynamic and efficient

next generation wireless systems. Then, the fundamental concepts and theoretical analyses of the dynamic

spectrum access have been presented. The various approaches of spectrum sharing in cognitive radio have

been reviewed. The security and economic issues have been discussed. From the practical perspective,

the reviews of simulation tools and testbed have been provided. The future trends and research directions

have been discussed. Also, the standardization activities related to cognitive radio have been summarized.

Page 52 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

51

REFERENCES

[1] Guest Editorial, KICS/IEEE Journal of Communications and Networks (JCN) Special Issue on “Cognitive radio: A path in the evolution

of public wireless networks”, December 2008.

[2] S. Haykin, “Cognitive radio: Brain-empowered wireless communications,” IEEE Journal on Selected Areas in Communications, vol.

23, no. 2, pp. 201–220, February 2005.

[3] Future Directions in Cognitive Radio Network Research, NSF Workshop Report, Eds. P. Steenkiste, D. Sicker, G. Minden, and D.

Raychaudhuri, March 9-10, 2009. [Online] Available at: http://www.cs.cmu.edu/∼prs/NSF CRN Report Final.pdf

[4] J. Mitola III and G. Q. Maguire Jr., “Cognitive radio: Making software radios more personal,” IEEE Personal Communications, vol.

6, no. 4, pp. 13–18, August 1999.

[5] J. Mitola III, Cognitive radio: An Integrated Agent Architecture for Software Defined Radio, Doctor of Technology Thesis, Royal Inst.

Technol. (KTH), Stockholm, Sweden, 2000.

[6] Federal Communications Commission, “Spectrum policy task force report.” ET Docket No. 02-135, 2002.

[7] Federal Communications Commission, “In the Matter of Unlicensed Operation in the TV Broadcast Bands,” FCC 08-260, November

2008.

[8] Federal Communications Commission, “Unlicensed operation in the TV broadcast bands and additional spectrum for unlicensed

devices below 900 MHz and in the 3 GHz band,” FCC 10-174, ET Docket No. 04-186 and 02-380, September 2010.

[9] I. F. Akyildiz, W.-Y. Lee, M. C. Vuran, and S. Mohanty, “NeXt generation/dynamic spectrum access/cognitive radio wireless networks:

A survey,” Computer Networks (Elsevier), vol. 50, no. 13, 15 September 2006, pp. 2127–2159.

[10] M. Buddhikot, “Understanding dynamic spectrum access: Models, taxonomy and challenges,” in Proc. of IEEE DySPAN 2007, Dublin,

Ireland, April 17-21, 2007.

[11] Y.-C. Liang, K.-C. Chen, G. Y. Li, and P. Mahonen, “Cognitive radio networking and communications: An overview,” IEEE

Transactions on Vehicular Technology, vol. 60, no. 7, pp. 3386–3407, September 2011.

[12] B. Wang and K. J. R. Liu, “Advances in cognitive radio networks: A survey,” IEEE Journal of Selected Topics in Signal Processing,

vol. 5, no. 1, pp. 5–23, February 2011.

[13] Q. Zhao and B. M. Sadler, “A survey of dynamic spectrum access,” IEEE Signal Processing Magazine, vol. 24, no. 3, pp. 79–89,

May 2007.

[14] A. Goldsmith, S. A. Jafar, I. Maric, and S. Srinivasa, “Breaking spectrum gridlock with cognitive radios: An information theoretic

perspective,” Proceedings of the IEEE, vol. 97, no. 5, pp. 894–914, May 2009

[15] E. Hossain, D. Niyato, and Z. Han, Dynamic Spectrum Access and Management in Cognitive Radio Networks. Cambridge University

Press, 2009.

[16] Cognitive Radio Communications and Networks: Principles and Practice. (Eds. Alexander M. Wyglinski, Maziar Nekovee, and Y.

Thomas Hou), Elsevier, December 2009.

[17] Cognitive Wireless Communication Networks. (Eds. E. Hossain and V. K. Bhargava), ISBN: 978-0-387-68830-5, Springer, 2007.

[18] Cognitive Radio Networks. (Eds. Yang Xiao and Fei Hu), 978-1-4200-6420-9, CRC Press, 2009.

[19] J. Mitola III, Cognitive Radio Architecture: The Engineering Foundations of Radio XML. ISBN: 978-0471742449, Wiley Press, 2006.

[20] B. A. Fette, Cognitive Radio Technology (2nd Edition). Academic Press, ISBN: 978-0123745354, 2009.

[21] L. E. Doyle, Essentials of Cognitive Radio. Cambridge University Press, 2009.

[22] E. Biglieri, A. J. Goldsmith, L. J. Greenstein, N. Mandayam, and H. V. Poor, Principles of Cognitive Radio, Cambridge University

Press, December 2012.

[23] K.-C. Chen and R. Prasad, Cognitive Radio Networks. John Wiley and Sons, 2009.

Page 53 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

52

[24] IEEE Journal on Selected Areas in Communications, Special Issue on “Cognitive Radio: Theory and Application”, vol. 26, no. 1,

January 2008.

[25] Proceedings of the IEEE, Special Issue on “Cognitive Radio - I”, vol. 97, no. 4, April 2009.

[26] Proceedings of the IEEE, Special Issue on “Cognitive Radio - II”, vol. 97, no. 5, May 2009.

[27] IEEE Journal on Selected Areas in Communications, Special Issue on “Advances in Cognitive Radio Networking and Communications

- I”, vol. 29, no. 2, February 2011.

[28] IEEE Journal on Selected Areas in Communications, Special Issue on “Advances in Cognitive Radio Networking and Communications

- II”, vol. 29, no. 4, April 2011.

[29] IEEE Transactions on Vehicular Technology, Special issue on the “Achievements and the Road Ahead: The First Decade of Cognitive

Radio”, vol. 59, no. 4, May 2010.

[30] IEEE Journal of Selected Topics in Signal Processing, Special Issue on “Signal Processing in Cooperative Cognitive Radio Systems”,

vol. 5, no. 1, January 2011.

[31] IEEE Wireless Communications Magazine, Special Issue on “Cognitive Radio Networks: A Practical Perspective”, August 2012.

[32] IEEE Vehicular Technology Magazine, Special Issue on “Applications of Cognitive Radio Networks”, June 2012.

[33] EURASIP Journal on Wireless Communications and Networking, Special Issue on “Ten Years of Cognitive Radio: State of the Art

and Perspectives”, November 2011.

[34] EURASIP Journal on Advances in Signal Processing, Special Issue on “Dynamic Spectrum Access for Wireless Networking”, August

2009.

[35] EURASIP Journal on Advances in Signal Processing, Special Issue on “Advanced Signal Processing for Cognitive Radio Networks”,

2010.

[36] Elsevier Computer Communications, Special Issue on “Cognitive Radio and Dynamic Spectrum Sharing Systems”, vol. 32, no. 18,

December 2009.

[37] R. Deng, J. Chen, X. Cao, Y. Zhang, S. Maharjan, and S. Gjessing, “Sensing-Performance Tradeoff in Cognitive Radio Enabled Smart

Grid,” IEEE Transactions on Smart Grid, vol. 4, no. 1, pp. 302-310, March 2013.

[38] J. Huang, H. Wang, Y. Qian, and C. Wang, “Priority-Based Traffic Scheduling and Utility Optimization for Cognitive Radio

Communication Infrastructure-Based Smart Grid,” IEEE Transactions on Smart Grid, vol. 4, no. 1, pp. 78-86, March 2013.

[39] Y. Zhang, R. Yu, M. Nekovee, Y. Liu, S. Xie, and S. Gjessing, “Cognitive machine-to-machine communications: visions and potentials

for the smart grid,” IEEE Network, vol. 26, no. 3, pp. 6-13, May-June 2012.

[40] S.-H. Wu, H.-L. Chao, C.-H. Ko, S.-R. Mo, C.-T. Jiang, T.-L. Li, C.-C. Cheng, and C.-F. Liang, “A cloud model and concept prototype

for cognitive radio networks,” IEEE Wireless Communications, vol. 19, no. 4, pp. 49-58, August 2012.

[41] A. Ghasemi and E. S. Sousa, “Spectrum sensing in cognitive radio networks: Requirements, challenges and design trade-offs,” IEEE

Communications Magazine, vol. 46, no. 4, pp. 32–39, April 2008.

[42] A. Ghasemi and E. S. Sousa, “Interference aggregation in spectrum-sensing cognitive wireless networks,” IEEE Journal of Selected

Topics in Signal Processing, vol. 2, no. 1, pp. 41–56, February 2008.

[43] T. Yucek and H. Arslan, “A survey of spectrum sensing algorithms for cognitive radio applications,” IEEE Communications Surveys

and Tutorials, vol. 11, no. 1, pp.116–130, 2009.

[44] Y. Zeng, Y.-C. Liang, A. T. Hoang, and R. Zhang, “A review on spectrum sensing for cognitive radio: Challenges and solutions,”

EURASIP Journal on Advances in Signal Processing, volume 2010, Article ID 381465, 2010.

[45] P. D. Sutton, K. E. Nolan, and L. E. Doyle, “Cyclostationary signatures in practical cognitive radio applications,” IEEE Journal on

Selected Areas in Communications, vol. 26, no. 1, pp. 13–24, January 2008.

Page 54 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

53

[46] A. F. Molisch, M. Shafi, and L. J. Greenstein, “Propagation issues for cognitive radio,” Proceedings of the IEEE, Special Issue on

“Cognitive Radio”, 97, 787–804 (2009).

[47] A. Ghasemi and E. S. Sousa, “Fundamental limits of spectrum-sharing in fading environments,” IEEE Transactions on Wireless

Communications, vol. 6, no. 2, pp. 649–658, February 2007.

[48] Y. Zeng and Y.-C. Liang, “Eigenvalue based spectrum sensing algorithms for cognitive radio,” IEEE Transactions Communications,

vol. 57, no. 6, pp. 1784–1793, 2009.

[49] A. Mariani, A. Giorgetti, and M. Chiani, “Effects of noise power estimation on energy detection for cognitive radio applications,”

IEEE Transactions on Communications, vol. 59, no. 12, pp. 3410-3420, December 2011.

[50] R. Tandra and A. Sahai, “SNR walls for signal detection,” IEEE Journal of Selected Topics in Signal Processing, vol. 2, no. 1, pp.

4–17, February 2008.

[51] S. Haykin, D. Thomson, and J. Reed, “Spectrum sensing for cognitive radio,” Proceedings of the IEEE, Special Issue on “Cognitive

Radio”, vol. 97, no. 5, pp. 849–877, May 2009.

[52] Y.-C. Liang, Y. Zeng, E. C. Y. Peh, and A. T. Hoang, “Sensing-throughput tradeoff for cognitive radio networks,” IEEE Transactions

on Wireless Communications, vol. 7, no. 4, pp. 1326–1337, April 2008.

[53] K. B. Letaief and W. Zhang, “Cooperative communications for cognitive radio networks,” Proceedings of the IEEE, vol. 97, no. 5,

pp. 878–893, May 2009.

[54] G. Ganesan and Y. Li, “Cooperative spectrum sensing in cognitive radio, part I: Two user networks,” IEEE Transactions on Wireless

Communications, vol. 6, no. 6, pp. 2204–2213, June 2007.

[55] G. Ganesan and Y. Li, “Cooperative spectrum sensing in cognitive radio, part II: Multiuser networks,” IEEE Transactions on Wireless

Communications, vol. 6, no. 6, pp. 2214–2222, June 2007.

[56] J. Unnikrishnan and V. V. Veeravalli, “Cooperative sensing for primary detection in cognitive radio,” IEEE Journal of Selected Topics

in Signal Processing, vol. 2, no. 1, pp. 18–27, February 2008.

[57] R. Zhang and Y.-C. Liang, “Exploiting multi-antennas for opportunistic spectrum sharing in cognitive radio networks,” IEEE Journal

of Selected Topics in Signal Processing, vol. 2, no. 1, pp. 88–102, February 2008.

[58] Z. Quan, S. Cui, and A. H. Sayed, “Optimal linear cooperation for spectrum sensing in cognitive radio networks,” IEEE Journal of

Selected Topics in Signal Processing, vol. 2, no. 1, pp. 28–40, February 2008.

[59] J. Meng, W. Yin, H. Li, E. Hossain, and Z. Han, “Collaborative spectrum sensing from sparse observations in cognitive radio

networks,” IEEE Journal on Selected Topics on Communications, Special Issue on “Advances in Cognitive Radio Networking and

Communications”, vol. 29, no. 2, pp. 327–337, February 2011.

[60] K. W. Choi, E. Hossain, and D. I. Kim, “Cooperative spectrum sensing under random geometric primary user network model,” IEEE

Transactions on Wireless Communications, vol. 10, no. 6, pp. 1932–1944, June 2011.

[61] L. Lu, X. Zhou, U. Onunkwo, and G. Y. Li, “Ten years of research in spectrum sensing and sharing in cognitive radio,” EURASIP

Journal on Wireless Communications and Networking, 2012:28, January 2012.

[62] G. L. Stuber, S. M. Almalfouh, and D. Sale, “Interference analysis of TV-band white space,” Proceedings of the IEEE, vol. 97, no.

4, pp. 741–754, April 2009.

[63] A. Rabbachin, Tony Q. S. Quek, H. Shin, and M. Z. Win, “Cognitive network interference,” IEEE Journal on Selected Areas in

Communications, vol. 29, no. 2, February 2011.

[64] D. A. Roberson, C. S. Hood, J. L. LoCicero, and J. T. MacDonald, “Spectral occupancy and interference studies in support of cognitive

radio technology deployment,” in Proc. of First IEEE Workshop on Networking Technologies for Software Defined Radio Networks,

pp. 26–35, September 2006.

Page 55 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

54

[65] D. Datla, A. M. Wyglinski, and G. J. Minden, “A spectrum surveying framework for dynamic spectrum access networks,” IEEE

Transactions on Vehicular Technology, vol. 58, no. 8, pp. 4158–4158, April 2009.

[66] C. Ghosh, S. Pagadarai, D. Agrawal, and A. M. Wyglinski, “A framework for statistical wireless spectrum occupancy modeling,”

IEEE Transactions on Wireless Communications, vol. 9, no. 1, pp. 38–44, January 2010.

[67] M. Wellens and P. Mahonen, “Lessons learned from an extensive spectrum occupancy measurement campaign and a stochastic duty

cycle model,” Mobile Networks and Applications (Springer), 2009.

[68] B. Canberk, I. F. Akyildiz, and S. Oktug, “Primary user activity modeling using first-difference filter clustering and correlation in

cognitive radio networks,” IEEE/ACM Transactions on Networking, vol. 19, no. 1, pp. 170–183, February 2011.

[69] T. Jiang, Z.-G. Cui, J.-W. Hu, and Y.-L. Hou, “A design of cognitive UWB pulse based on APSWF,” in Proc. of Asia-Pacific Symposium

on Electromagnetic Compatibility (APEMC 2008), pp. 706–709, 19-23 May 2008.

[70] W. Zhang, H. Shen, Z. Bai, and K. S. Kwak, “A novel modulation waveform on ultra-wideband based cognitive radio systems,” in

Proc. of 2nd International Conference on Cognitive Radio Oriented Wireless Networks and Communications (CrownCom 2007), pp.

306–311, 1-3 August 2007.

[71] F. Wei, P. Xia, Z. Yang, and F. Tian, “Decentralized waveform design for MIMO cognitive radio under interference temperature

constraint,” in Proc. of 2011 Second International Conference on Networking and Distributed Computing (ICNDC), pp.159–162,

21-24 September 2011.

[72] L.-L. Zhou, H.-B. Zhu, and N.-T. Zhang, “Iterative solution to the notched waveform design in cognitive ultra-wideband radio system,”

Progress In Electromagnetic Research, vol. 75, pp. 271–284, 2007.

[73] Z. Tian, G. Leus, and V. Lottici, “Joint dynamic resource allocation and waveform adaptation in cognitive radio networks,” in Proc.

of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2008), pp. 5368–5371, March 31 2008-April

4 2008.

[74] V. Chakravarthy, X. Li, Z. Wu, M. Temple, F. Garber, R. Kannan, and A. Vasilakos, “Novel overlay/underlay cognitive radio waveforms

using SD-SMSE framework to enhance spectrum efficiency- part i: Theoretical framework and analysis in AWGN channel,” IEEE

Transactions on Communications, vol. 57, no. 12, pp. 3794–3804, December 2009.

[75] S. Pagadarai, A. Kliks, H. Bogucka, and A. M. Wyglinski, “On non-contiguous multicarrier waveforms for spectrally opportunistic

cognitive radio systems,” in Proc. of 2010 International Waveform Diversity and Design Conference (WDD), pp. 177–181, 8-13 August

2010.

[76] Z. Hu, N. Guo, and R. Qiu, “Wideband waveform design for relay cognitive network,” in Proc. of IEEE Military Communications

Conference (Milcom 2010), pp. 749–754, October 31-November 3, 2010.

[77] Z. Hu, N. Guo, and R. Qiu, “Wideband waveform optimization for multiple input single output cognitive radio with practical

considerations,” in Proc. of IEEE Military Communications Conference (Milcom 2010), pp. 1227–1232, October 31-November 3,

2010.

[78] Z. Kollar and P. Horvath, “Physical layer considerations for cognitive radio: Modulation techniques,” in Proc. of 2011 IEEE 73rd

Vehicular Technology Conference (VTC Spring), pp. 1–5, 15-18 May 2011.

[79] Y. Chen, M.-S. Alouini, and L. Tang, “Performance analysis of adaptive modulation for cognitive radios with opportunistic access,”

in Proc. of 2011 IEEE International Conference on Communications (ICC), pp. 1–5, 5-9 June 2011.

[80] M. Zamanian, A. A. Tadaion, and M. T. Sadeghi, “Modulation classification of linearly modulated signals in a cognitive radio

network using constellation shape,” in Proc. of 2011 7th International Workshop on Systems, Signal Processing and their Applications

(WOSSPA), pp. 13–16, 9-11 May 2011.

[81] M. R. Khanzadi, K. Haghighi, A. Panahi, and T. Eriksson, “A novel cognitive modulation method considering the performance of

Page 56 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

55

primary user,” in Proc. of 2010 6th Conference on Wireless Advanced (WiAD), pp. 1–6, 27-29 June 2010.

[82] Z. Ye, G. Memik, and J. Grosspietsch, “Digital modulation classification using temporal waveform features for cognitive radios,”

Proc. of IEEE Int. Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC 2007), pp. 1–5, 3-7 September 2007.

[83] I. Budiarjo, H. Nikookar, and L. Ligthart, “Cognitive radio modulation techniques,” IEEE Signal Processing Magazine, vol. 25, no.

6, pp. 24–34, November 2008.

[84] R. Rajbanshi, Q. Chen, A. M. Wyglinski, G. J. Minden, and J. B. Evans, “Quantitative comparison of agile modulation techniques for

cognitive radio transceivers,” in Proc. of 4th IEEE Consumer Communications and Networking Conference (CCNC 2007), pp. 1144–

1148, 11–13 January 2007.

[85] P. Kryszkiewicz, H. Bogucka, and A. M. Wyglinski, “Protection of primary users in dynamically varying radio environment: Practical

solutions and challenges,” EURASIP Journal on Wireless Communications and Networking, 2012:23.

[86] D. Qu, Z. Wang, and T. Jiang, “Extended active interference cancellation for sidelobe suppression in cognitive radio OFDM systems

with cyclic prefix,” IEEE Transactions on Vehicular Technology, vol. 59, no. 4, pp.1689–1695, May 2010.

[87] A. Ghasemi and E. S. Sousa, “Interference aggregation in spectrum-sensing cognitive wireless networks,” IEEE Journal of Selected

Topics in Signal Processing, vol. 2, no. 1, pp. 41–56, February 2008.

[88] R. Rajbanshi, Q. Chen, A. M. Wyglinski, J. B. Evans, and G. J. Minden, “Comparative study of frequency agile data transmission

schemes for cognitive radio transceivers,” in Proc. of 1st Int. Wksp. on Technol. and Policy for Accessing Spectrum, Boston, MA,

USA, August 2006.

[89] A. Batra, S. Lingam, and J. Balakrishnan, “Multi-band OFDM: A cognitive radio for UWB,” in Proc. of IEEE International Symposium

on Circuits and Systems (ISCAS 2006), 21-24 May 2006.

[90] S.-M. Cheng, W. C. Ao, and K.-C. Chen, “Efficiency of a cognitive radio link with opportunistic interference mitigation,” IEEE

Transactions on Wireless Communications, vol. 10, no. 6, pp. 1715–1720, June 2011.

[91] H. Bogucka, A. M. Wyglinski, S. Pagadarai, A. Kliks, “Spectrally agile multicarrier waveforms for opportunistic wireless access,”

IEEE Communications Magazine, vol. 49, no. 6, pp. 108–115, June 2011.

[92] M. M. Ramon, T. Atwood, S. Barbin, and C. G. Christodoulou, “Signal classification with an SVM-FFT approach for feature extraction

in cognitive radio,” in Proc. of SBMO/IEEE MTT-S International Microwave & Optoelectronics Conference (IMOC 2009).

[93] M. G. Khoshkholgh, K. Navaie, and H. Yanikomeroglu, “Access strategies for spectrum sharing in fading environment: Overlay,

underlay and mixed,” IEEE Transactions on Mobile Computing, vol. 9, no. 12, pp. 1780–1793, December 2010.

[94] R. A. Tannious and A. Nosratinia, “Cognitive radio protocols based on exploiting hybrid ARQ retransmissions,” IEEE Transactions

on Wireless Communications, vol. 9, no. 9, pp. 2833–2841, September 2010.

[95] Y. Xing, R. Chandramouli, S. Mangold, and S. Sankar N, “Dynamic spectrum access in open spectrum wireless networks,” IEEE

Journal on Selected Areas in Communications, vol. 24, no. 3, pp. 626-637, March 2006.

[96] L. B. Le and E. Hossain, “Resource allocation for spectrum underlay in cognitive wireless networks,” IEEE Transactions on Wireless

Communications, vol. 7, no. 12, December 2008, pp. 5306–5315.

[97] Y.-C. Liang, Y. Zeng, E.C.Y. Peh, and A. T. Hoang, “Sensing-throughput tradeoff for cognitive radio networks,” IEEE Transactions

on Wireless Communications, vol. 7, no. 4, pp. 1326–1337, April 2008.

[98] D. I. Kim, L. B. Le, and E. Hossain, “Joint rate and power allocation for cognitive radios in dynamic spectrum access environment,”

IEEE Transactions on Wireless Communications, vol. 7, no. 12 - part 2, pp. 5517–5527, December 2008.

[99] R. Zhang and Y.-C. Liang, “Exploiting multi-antennas for opportunistic spectrum sharing in cognitive radio networks,” IEEE Journal

of Selected Topics in Signal Processing, vol. 2, no.1, pp. 88–102, February 2008.

Page 57 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

56

[100] L. Zhang, Y.-C. Liang, and Y. Xin, “Joint beamforming and power allocation for multiple access channels in cognitive radio networks,”

IEEE Journal on Selected Areas in Communications, vol. 26, no.1, pp. 38–51, January 2008.

[101] R. Zhang, Y.-C. Liang and S. Cui, “Dynamic resource allocation in cognitive radio networks,” IEEE Signal Processing Magazine,

vol. 27, no. 3, pp. 102–114, May 2010.

[102] E. Hossain, L. Le, N. Devroye, and M. Vu, “Cognitive radio: From theory to practical network engineering,” invited chapter in

Advances in Wireless Communications, (Ed. V. Tarokh), Springer, 2009.

[103] S. C. Jha, M. M. Rashid, V. K. Bhargava, and C. L. Despins, “Medium access control in distributed cognitive radio networks,” IEEE

Wireless Communications, August 2011, pp. 41–51.

[104] H. Kim and K. G. Shin, “Efficient discovery of Spectrum opportunities with MAC-layer sensing in cognitive radio networks,” IEEE

Transactions on Mobile Computing, vol. 7, no. 5, pp. 533–545, May 2008.

[105] Q. Zhao, L. Tong, A. Swami, and Y. Chen, “Decentralized cognitive MAC for opportunistic spectrum access in ad hoc networks: A

POMDP framework,” IEEE Journal on Selected Areas in Communications, vol. 25, no. 3, pp. 589–600, April 2007.

[106] J. Jia, Q. Zhang, and X. Shen, “HC-MAC: A hardware-constrained cognitive MAC for efficient spectrum management,” IEEE Journal

on Selected Areas in Communications, vol. 26, no. 1, pp. 106–117, January 2008.

[107] S. Wang and J. Zhang, “Opportunistic Spectrum Scheduling by Jointly Exploiting Channel Correlation and PU Traffic Memory,” IEEE

Journal on Selected Areas in Communications, vol. 31, no. 3, pp. 394-405, March 2013.

[108] Y. Chen, Q. Zhao, and A. Swami, “Joint design and separation principle for opportunistic spectrum access in the presence of sensing

errors,” IEEE Transactions on Information Theory, vol. 54, no. 5, pp. 2053–2071, May 2008.

[109] Y. Xing, R. Chandramouli, S. Mangold, and S. Sankar N, “Dynamic spectrum access in open spectrum wireless networks,” IEEE

Journal on Selected Areas in Communications, vol. 24, no. 3, pp. 626–637, March 2006.

[110] Y. Xing, C. N. Mathur, M. A. Haleem, R. Chandramouli, and K. P. Subbalakshmi, “Dynamic spectrum access with QoS and interference

temperature constraints,” IEEE Transactions on Mobile Computing, vol. 6, no. 4, pp. 423–433, April 2007.

[111] H. Su and X. Zhang, “Cross-layer based opportunistic MAC protocols for QoS provisioning over cognitive radio wireless networks,”

IEEE Journal on Selected Areas in Communications, vol. 26, no. 1, pp. 118–129, January 2008.

[112] H. Jiang, L. Lai, R. Fan, and H. V. Poor, “Optimal selection of channel sensing order in cognitive radio,” IEEE Transactions on

Wireless Communications, vol. 8, no. 1, pp. 297–303, January 2009.

[113] H. Su and X. Zhang, “Cross-layer based opportunistic MAC protocols for QoS provisionings over cognitive radio wireless networks,”

IEEE Journal on Selected Areas in Communications, vol. 26, no. 1, pp. 118-129, Jan. 2008.

[114] B. Jalaeian and M. Motani, “Location Aware CR-MAC: A multi-channel cross layered PHY-MAC protocol for cognitive radio ad hoc

networks,” in Proceedings of IEEE Malaysia International Conference on Communications (MICC), pp. 348-353, 15-17 Dec. 2009.

[115] W. S. Jeon, D. H. Lee, and D. G. Jeong, “Collaborative Sensing Management for Cognitive Radio Networks with Reporting Overhead,”

IEEE Transactions on Wireless Communications, vol. 12, no. 2, pp. 595-605, February 2013.

[116] X. Kang, Y.-C. Liang, A. Nallanathan, H. K. Garg, and R. Zhang, “Optimal power allocation for fading channels in cognitive radio

networks: Ergodic capacity and outage capacity,” IEEE Transactions on Wireless Communications, vol. 8, no. 2, pp. 940–950, February

2009.

[117] G. Bansal, M. J. Hossain, and V. K. Bhargava, “Optimal and suboptimal power allocation schemes for OFDM-based cognitive radio

systems,” IEEE Trans. Wireless Commun., 7(11): 4710–4718, November 2008.

[118] G. Bansal, M. J. Hossain, and V. K. Bhargava, “Adaptive power loading for OFDM-based cognitive radio systems with statistical

interference constraint,” IEEE Trans. Wireless Commun., vol. 10, no. 9, pp. 2786–2791, September 2011.

Page 58 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

57

[119] D. Bharadia, G. Bansal, P. Kaligineedi, and V. K. Bhargava, “Relay and power allocation schemes for OFDM-based cognitive radio

systems,” IEEE Trans. Wireless Commun., vol. 10, no. 9, pp. 2812–2817, September 2011.

[120] Z. Hasan, G. Bansal, E. Hossain, and V. K. Bhargava, “Energy-efficient power allocation in OFDM-based cognitive radio systems: A

risk-return model,” IEEE Trans. Wireless Commun., 8(12): 6078–6088, December 2009.

[121] P. Phunchongharn and E. Hossain, “Distributed robust scheduling and power control for cognitive spatial-reuse TDMA networks,”

IEEE Journal on Selected Areas in Communications, vol. 30, no. 10, pp. 1934-1946, November 2012.

[122] M. M. Rashid, M. J. Hossain, E. Hossain, and V. K. Bhargava, “Opportunistic spectrum scheduling for multiuser cognitive radio: A

queueing analysis,” IEEE Trans. Wireless Commun., 8(10): 5259–5269, October 2009.

[123] I. F. Akyildiz, Won-Yeol Lee, M. C. Vuran, and S. Mohanty, “A survey on spectrum management in cognitive radio networks,” IEEE


[124] L.-C. Wang, C.-W. Wang, and K.-T. Feng, “A queueing-theoretical framework for QoS-enhanced spectrum management in cognitive

radio networks,” IEEE Wireless Communications, vol. 18, no. 6, pp. 18–26, December 2011.

[125] R. Zhou, X. Li, V. Chakravarthy, and Z. Wu, “Spectrum mobility demonstration of SMSE based overlay cognitive radio via software

defined radio,” in Proc. of IEEE Symposium on New Frontiers in Dynamic Spectrum Access Networks (DySPAN), pp. 668–669, May

2011.

[126] R. W. Thomas, D. H. Friend, L. A. Dasilva, and A. B. Mackenzie, “Cognitive networks: Adaptation and learning to achieve end-to-end

performance objectives,” IEEE Communications Magazine, vol. 44, no. 12, pp. 51–57, December 2006.

[127] N. Nie and C. Comaniciu, “Adaptive channel allocation spectrum etiquette for cognitive radio networks,” in Proc. of First IEEE

International Symposium on Dynamic Spectrum Access Networks (DySPAN05), pp. 269–278, November 2005.

[128] J.-S. Pang and G. Scutari, “Joint Sensing and Power Allocation in Nonconvex Cognitive Radio Games: Quasi-Nash Equilibria,” IEEE

Transactions on Signal Processing, vol. 61, no. 9, pp. 2366-2382, May 2013.

[129] A. He, K. K. Bae, T. Newman, J. Gaeddert, K. Kim, R. Menon, L. Morales-Tirado, J. Neel, Y. Zhao, J. Reed, and W. Tranter, “A

survey of artificial intelligence for cognitive radios,” IEEE Transactions on Vehicular Technology, vol. 59, no. 4, pp. 1578–1592, May

2010.

[130] Y. Xing and R. Chandramouli, “Human behavior inspired cognitive radio network design,” IEEE Communications Magazine, vol. 46,

no. 12, pp. 122–127, December 2008.

[131] C. Clancy, J. Hecker, E. Stuntebeck, and T. OShea, “Applications of machine learning to cognitive radio networks,” IEEE Wireless

Communications, vol. 14, no. 4, pp. 47–52.

[132] A. G.-Serrano and L. Giupponi, “Distributed Q-learning for aggregated interference control in cognitive radio networks,” IEEE

Transactions on Vehicular Technology, vol. 59, no. 4, pp. 1823–1834, May 2010.

[133] Z. Han, R. Zheng, and H. Poor, “Repeated auctions with bayesian nonparametric learning for spectrum access in cognitive radio

networks,” IEEE Transactions on Wireless Communications, vol. 10, no. 3, pp. 890–900, Mar. 2011.

[134] T. Clancy, A. Khawar, and T. Newman, “Robust signal classification using unsupervised learning,” IEEE Transactions on Wireless

Communications, vol. 10, no. 4, pp. 1289–1299, April 2011.

[135] M. Maskery, V. Krishnamurthy, and Q. Zhao, “Decentralized dynamic spectrum access for cognitive radios: Cooperative design of a

non-cooperative game,” IEEE Transactions on Communications, vol. 57, no. 2, pp. 459–469, February 2009.

[136] M. van der Schaar and F. Fu, “Spectrum access games and strategic learning in cognitive radio networks for delay-critical applications,”

Proceedings of the IEEE, vol. 97, no. 4, pp. 720–740, April 2009.

[137] N. Baldo, B. Tamma, B. Manojt, R. Rao, and M. Zorzi, “A neural network based cognitive controller for dynamic channel selection,”

in Proc. of IEEE International Conference on Communications (ICC09), June 2009, pp. 1–5.

Page 59 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

58

[138] V. Tumuluru, P. Wang, and D. Niyato, “A neural network based spectrum prediction scheme for cognitive radio,” in Proc. of IEEE

International Conference on Communications (ICC’10), May 2010, pp. 1–5.

[139] A. Anandkumar, N. Michael, A. K. Tang, and A. Swami, “Distributed algorithms for learning and cognitive medium access with

logarithmic regret,” IEEE JSAC on Advances in Cognitive Radio Networking and Communications, vol. 29, no. 4, pp. 781–745, April

2011.

[140] K. W. Choi and E. Hossain, “Opportunistic access to spectrum holes between packet bursts: A learning-based approach,” IEEE

Transactions on Wireless Communications, vol. 10, no. 8, pp. 2497–2509, August 2011.

[141] I. F. Akyildiz, W.-Y. Lee, and K. R. Chowdhury, “CRAHNs: Cognitive radio ad hoc networks,” Ad Hoc Networks (Elsevier), vol. 7,

no. 5, July 2009, pp. 810–836.

[142] Z. Yang, G. Cheng, W. Liu, W. Yuan, W. Cheng, “Local coordination based routing and spectrum assignment in multi-hop cognitive

radio networks,” ACM MONET, vol. 13, pp. 67–81, April 2008.

[143] Y. T. Hou, Y. Shi, and H. D. Sherali, “Spectrum sharing for multi-hop networking with cognitive radios,” IEEE Journal on Selected

Areas in Communications, vol. 26, no. 1, pp. 146–155, January 2008.

[144] R. Urgaonkar and M. J. Neely, “Opportunistic scheduling with reliability guarantees in cognitive radio networks,” IEEE Transactions

on Mobile Computing, vol. 8, no. 6, pp. 766–777, June 2009.

[145] D. Xue and E. Ekici, “Cross-Layer Scheduling for Cooperative Multi-Hop Cognitive Radio Networks,” IEEE Journal on Selected

Areas in Communications, vol. 31, no. 3, pp. 534-543, March 2013.

[146] S. H. Song, M. O. Hasna, and K. B. Letaief, “Prior Zero Forcing for Cognitive Relaying,” IEEE Transactions on Wireless

Communications, vol. 12, no. 2, pp. 938-947, February 2013.

[147] J.-C. Liang and J.-C. Chen, “Resource Allocation in Cognitive Radio Relay Networks,” IEEE Journal on Selected Areas in

Communications, vol. 31, no. 3, pp. 476-488, March 2013.

[148] Y. Wu and D. H. K. Tsang, “Dynamic rate allocation, routing and spectrum sharing for multi-hop cognitive radio networks,” IEEE

International Conference on Communications (ICC’09) Workshops, pp. 1–6, June 2009.

[149] T. C. Clancy and N. Goergen, “Security in cognitive radio networks: Threats and mitigation,” in Proc. of 3rd International Conference

on Cognitive Radio Oriented Wireless Networks and Communications (CrownCom 2008), pp. 1–8, May 2008.

[150] R. Chen, J.-M. Park, and J. H. Reed, “Defense against primary user emulation attacks in cognitive radio networks,” IEEE Journal on

Selected Areas in Communications, vol. 26, no. 1, pp. 25–37, January 2008.

[151] R. Chen, J.-M. Park, Y. T. Hou, and J. H. Reed, “Toward secure distributed spectrum sensing in cognitive radio networks,” IEEE


[152] S. Zheng, P.-Y. Kam, Y.-C. Liang, and Y. Zeng, “Spectrum sensing for digital primary signals in cognitive radio: A Bayesian approach

for maximizing spectrum utilization,” IEEE Transactions on Wireless Communications, vol. 12, no. 4, pp. 1774-1782, April 2013.

[153] H. Li and Z. Han, “Dogfight in spectrum: Combating primary user emulation attacks in cognitive radio systems, part I: Known channel

statistics,” IEEE Transactions on Wireless Communications, vol. 9, no. 11, pp. 3566–3577, November 2010

[154] H. Li and Z. Han, “Dogfight in spectrum: Combating primary user emulation attacks in cognitive radio systems, part II: Unknown

channel statistics,” IEEE Transactions on Wireless Communications, vol. 10, no. 1, pp. 274–283, January 2011.

[155] H. Li and Z. Han, “Catch me if you can: An abnormality detection approach for collaborative spectrum sensing in cognitive radio

networks,” IEEE Transactions on Wireless Communications, vol. 9, no.11, pp. 3554–3565, November 2010.

[156] D. Niyato and E. Hossain, “Spectrum trading in cognitive radio networks: A market-equilibrium-based approach,” IEEE Wireless

Communications Magazine, vol. 15, no. 6, pp. 71–80, December 2008.

Page 60 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

59

[157] J. Huang, R. Berry, and M. Honig, “Auction-based spectrum sharing,” Mobile Networks and Applications, vol. 11, no. 3, pp. 405–418,

June 2006.

[158] J. Sun, E. Modiano, and L. Zheng, “Wireless channel allocation using an auction algorithm,”, IEEE Journal on Selected Areas in

Communications, vol. 24, no. 5, pp. 1085–1096, May 2006.

[159] Y. Chen, Y. Wu, B. Wang, and K. J. R. Liu, “Spectrum auction games for multimedia streaming over cognitive radio networks,” IEEE

Transactions on Communications, vol. 58, no. 8, pp. 2381–2390, August 2010.

[160] G. S. Kasbekar and S. Sarkar, “Spectrum auction framework for access allocation in cognitive radio networks,” IEEE/ACM Transactions

on Networking, vol. 18, no. 6, pp. 1841–1854, December 2010.

[161] S. Sodagari, A. Attar, and S. G. Bilen, “On a truthful mechanism for expiring spectrum sharing in cognitive radio networks,” IEEE

Journal on Selected Areas in Communications, vol. 29, no. 4, pp. 856–865, April 2011.

[162] Z. Han, R. Zheng, and H. V. Poor, “Repeated auctions with Bayesian nonparametric learning for spectrum access in cognitive radio

networks,” IEEE Transactions on Wireless Communications, vol. 10, no. 3, pp. 890–900, March 2011.

[163] F. Fu and M. van der Schaar, “Learning to compete for resources in wireless stochastic games,” IEEE Transactions on Vehicular

Technology, vol. 58, no. 4, pp. 1904–1919, May 2009.

[164] D. Niyato and E. Hossain, “Competitive pricing for spectrum sharing in cognitive radio networks: Dynamic game, inefficiency of

Nash equilibrium, and collusion,” IEEE Journal on Selected Areas in Communications, vol. 26, no. 1, pp. 192–202, January 2008

[165] F. Wang, M. Krunz, and S. Cui, “Price-based spectrum management in cognitive radio networks,” IEEE Journal on Selected Topics

in Signal Processing, vol. 2, no. 1, pp. 74–87, February 2008.

[166] Z. Ji and K. J. R. Liu, “Dynamic spectrum sharing: A game theoretical overview,” IEEE Communications Magazine, vol. 45, no. 5,

May 2007.

[167] D. Niyato and E. Hossain, “Competitive spectrum sharing in cognitive radio networks: A dynamic game approach,” IEEE Transactions

on Wireless Communications, vol. 7, no. 7, July 2008, pp. 2651–2660.

[168] D. Niyato, E. Hossain, and Z. Han, “Dynamics of multiple-seller and multiple-buyer spectrum trading in cognitive radio networks: A

game theoretic modeling approach,” IEEE Transactions on Mobile Computing, vol. 8, no. 8, August 2009, pp. 1009–1022.

[169] Z. Ji and K. J. R. Liu, “Multi-stage pricing game for collusion-resistant dynamic spectrum allocation,” IEEE Journal on Selected

Areas in Communications, vol. 26, no. 1, pp. 182–191, January 2008.

[170] L. Yang, H. Kim, J. Zhang, M. Chiang, and C. W. Tan, “Pricing-based decentralized spectrum access control in cognitive radio

networks,” IEEE/ACM Transactions on Networking, vol. 21, no. 2, pp. 522-535, April 2013.

[171] S. M. M. Toroujeni, S. M. S. Sadough, and S. A. G. Ghorashi, “Spectrum Leasing for OFDM-Based Cognitive Radio Networks,”

IEEE Transactions on Vehicular Technology, vol. 62, no. 5, pp. 2131-2139, Jun 2013.

[172] A. Pereira, H. Fontes, and A. Gameiro, “Simulation platform for opportunistic radio systems,” in Proc. of IEEE International Symposium

on Personal, Indoor and Mobile Radio Communications (PIMRC), September 2006.

[173] M. Zhan, P. Ren, and M. Gong, “An open software simulation platform for cognitive radio,” in Proc. of International Conference on

Wireless Communications Networking and Mobile Computing (WiCOM), September 2010.

[174] A. Sibille, “Efficient generation of spatially and frequency correlated random values for cognitive radio network simulators,” IEEE

Transactions on Vehicular Technology, vol. 59, no. 3, pp. 1121–1128, March 2010.

[175] A. A. Abidi, “The path to the software-defined radio receiver,” IEEE Journal of Solid-State Circuits, vol. 42, no. 5, pp. 954–966, May

2007.

[176] J. Mitola, “The software radio architecture,” IEEE Communications Magazine, vol. 33, no. 5, pp. 26–38, May 1995.

Page 61 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

60

[177] J. Mitola III, “Software radio architecture: A mathematical perspective,” IEEE Journal on Selected Areas in Communications, vol.

17, no. 4, pp. 514–538, April 1999.

[178] S. M. Mishra, D. Cabric, D. Chang, D. Willkomm, B. van Schewick, S. Wolisz, and B. W. Brodersen, “A real time cognitive radio

testbed for physical and link layer experiments,” in Proc. of First IEEE International Symposium on New Frontiers in Dynamic

Spectrum Access Networks (DySPAN05), pp. 562–567, 8-11 November 2005.

[179] D. Cabric, I. D. O’Donnell, M. S.-W. Chen, and R. W. Brodersen, “Spectrum sharing radios,” IEEE Circuits and Systems Magazine,

vol. 6, no. 2, pp. 30–45, 2006.

[180] R. Bagheri, A. Mirzaei, S. Chehrazi, M. E. Heidari, M. Lee, M. Mikhemar, W. Tang, and A. A. Abidi, “An 800 MHz - 6 GHz

software-defined wireless receiver in 90-nm CMOS,” IEEE Journal of Solid-State Circuits, vol. 41, no. 12, pp. 2860–2876, December

2006.

[181] http://www.cs.cmu.edu/∼emulator

[182] http://www.orbit-lab.org

[183] http://prisms.cs.umass.edu/dome/umassdieselnet

[184] http://www.citysense.net/

[185] http://www.winlab.rutgers.edu/docs/focus/WiNC2R.html

[186] G. J. Minden, J. B. Evans, L. Searl, D. DePardo, V. R. Petty, R. Rajbanshi, J. Guffey, Q. Chen, T. Newman, F. Weidling, D. Datla,

B. Barker, M. Peck, B. Cordill, A. M. Wyglinski, and A. Agah, “KUAR: A flexible software-defined radio development platform,” in

Proc. of Second IEEE Symposium on New Frontiers in Dynamic Spectrum Access Networks, Dublin, Ireland, April 2007.

[187] P. Marshall, “Extending the reach of cognitive radio,” Proceedings of the IEEE, vol. 97, no. 4, April 2009, pp. 612–625.

[188] GNU Radio Project, http://www.gnu.org/software/gnuradio

[189] WARP: Wireless Open Access Research Platform, 2009. http://warp.rice.edu

[190] J. N. Laneman, D. N. C. Tse, and G. W. Wornell, “Cooperative diversity in wireless networks: Efficient protocols and outage behavior,”

IEEE Trans. Info. Theory, vol. 50, no. 12, pp. 3062–3080, December 2004.

[191] Q. Zhang, J. Jia, and J. Zhang, “Cooperative relay to improve diversity in cognitive radio networks,” IEEE Communications Magazine,

vol. 47, no. 2, pp. 111–117, February 2009.

[192] H. Min, S. Lee, K. Kwak, and D. Hong, “Effect of multiple antennas at the source on outage probability for amplify-and-forward

relaying systems,” IEEE Trans. on Wireless Commun., vol. 8, no. 2, pp. 633–637, February 2009.

[193] K. B. Letaief and W. Zhang, “Cooperative communications for cognitive radio networks,” Proceedings of the IEEE, vol. 97, no. 5,

pp. 878 –893, May 2009.

[194] J. Jia, J. Zhang, and Q. Zhang, “Cooperative relay for cognitive radio networks,” IEEE INFOCOM 2009, pp. 2304–2312, April 2009.

[195] X. Gong et al., “A cooperative relay scheme for secondary communication in cognitive radio networks,” IEEE Global Telecommuni-

cations Conference (GLOBECOM 2008), pp. 1–6. December 2008.

[196] G. Zhao et al., “Spatial spectrum holes for cognitive radio with relay-assisted directional transmission,” IEEE Transactions on Wireless

Communications, vol. 8, no. 10, pp. 5270 –5279, October, 2009.

[197] J. Jia, J. Zhang, and Q. Zhang, “Relay-assisted routing in cognitive radio networks,” IEEE International Conference on Communications

(ICC’09), pp. 1–5, June 2009.

[198] Y. Zou et al., “A cognitive transmission scheme with the best relay selection in cognitive radio networks,” IEEE Global

Telecommunications Conference (GLOBECOM 2010), pp. 1–5, December 2010.

[199] L. Giupponi and C. Ibars, “Bayesian potential games to model cooperation for cognitive radios with incomplete information,” in IEEE

International Conference on Communications (ICC’09), pp. 1–6, June 2009.

Page 62 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

61

[200] Y. Zou et al., “An adaptive cooperation diversity scheme with best-relay selection in cognitive radio networks,” IEEE Transactions

on Signal Processing, vol. 58, no. 10, pp. 5438–5445, October 2010.

[201] Y. Li, R. Zhang, C. Wang, A. Daneshmand, and X. You, “A distributed relay selection algorithm for cognitive radio ad-hoc networks,”

2010 International Conference on Network and Service Management (CNSM), pp. 439–442, October 2010.

[202] J. Zhu, J. Huang, and W. Zhang, “Optimal one-dimensional relay placement in cognitive radio networks,” 2010 IEEE International

Conference on Wireless Communications and Signal Processing (WCSP), pp. 1–6, October 2010.

[203] A. Rao, H. Ma, M.-S. Alouini, and Y. Chen, “Impact of Primary User Traffic on Adaptive Transmission for Cognitive Radio with

Partial Relay Selection,” IEEE Transactions on Wireless Communications, vol. 12, no. 3, pp. 1162-1172, March 2013.

[204] V. N. Q. Bao, T. Q. Duong, D. Benevides da Costa, G. C. Alexandropoulos, and A. Nallanathan, “Cognitive Amplify-and-Forward

Relaying with Best Relay Selection in Non-Identical Rayleigh Fading,” IEEE Communications Letters, vol. 17, no. 3, pp. 475-478,

March 2013.

[205] Y. Li and A. Nosratinia, “Spectrum Sharing with Distributed Relay Selection and Clustering,” IEEE Transactions on Communications,

vol. 61, no. 1, pp. 53-62, January 2013.

[206] Y. Zou, Y.-D. Yao, and B. Zheng, “Cognitive transmissions with multiple relays in cognitive radio networks,” IEEE Transactions on

Wireless Communications, vol. 10, no. 2, pp. 648–659, February 2011.

[207] X. Qiao, Z. Tan, S. Xu, and J. Li, “Combined power allocation in cognitive radio-based relay-assisted networks,” IEEE International

Conference on Communications Workshops (ICC 2010), pp. 1–5, May 2010.

[208] Z. Liu, Y. Xu, D. Zhang, and S. Guan, “An efficient power allocation algorithm for relay assisted cognitive radio network,” 2010

International Conference on Wireless Communications and Signal Processing (WCSP), October 2010, pp. 1–5.

[209] O. Oyman, “Opportunistic scheduling and spectrum reuse in relay-based cellular OFDMA networks,” in IEEE Global Telecommu-

nications Conference (GLOBECOM ’07), pp. 3699–3703, November 2007.

[210] R. Wang, V. K. N. Lau, and Y. Cui, “Decentralized fair scheduling in two-hop relay-assisted cognitive OFDMA systems,” in IEEE

Journal on Selected Topics in Signal Processing, vol. 5, no. 1, pp. 171–181, February 2011.

[211] K. Lee and A. Yener, “Outage performance of cognitive wireless relay networks,” in IEEE Global Telecommunications Conference

(GLOBECOM ’06), pp. 1–5., December 2006.

[212] Y. Zou, Y.-D. Yao, and B. Zheng, “Outage probability analysis of cognitive transmissions: Impact of spectrum sensing overhead,” IEEE

Transactions on Wireless Communications, vol. 9, no. 8, pp. 2676–2688, August 2010.

[213] I. F. Akyildiz, D. M. G.-Estevenz, and E. C. Reyes, “The evolution to 4G cellular system: LTE-Advanced,” Physical Communication

(Elsevier), 3 (2010), pp. 217–244.

[214] J. Wang, M. Ghosh, and K. Challapali, “Emerging cognitive radio applications: A survey,” IEEE Communications Magazine, vol. 49,

no. 3, pp. 74–81, March 2011.

[215] W. Lehr and N. Jesuale, “Public safety radios must pool spectrum,” IEEE Communications Magazine, vol. 47, no. 3, pp. 103–109,

March 2009.

[216] E. Leu, B. L. Mark, and M. A. McHenry, “A framework for cognitive WiMAX with frequency agility,” Proceedings of the IEEE, vol.

97, no. 4, pp. 755–773, April 2009.

[217] M. Buddhikot, I. Kennedy, F. Mullany, and H. Viswanathan, “Ultra-broadband femtocells via opportunistic reuse of multi-operator

and multi-service spectrum,” Bell Labs Technical Journal (BLTJ), Special Issue on 4G Networks, February 2009.

[218] G. Gur, S. Bayhan, and F. Alagoz, “Cognitive femtocell networks: An overlay architecture for localized dynamic spectrum access,”

IEEE Wireless Communications, vol. 17, no. 4, pp. 62–70, August 2010.

Page 63 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

62

[219] T. Chen, H. Zhang, G. M. Maggio, and I. Chlamtac, “CogMesh: A cluster-based cognitive radio network,” in Proc. of 2nd IEEE

International Symposium on Dynamic Spectrum Access Networks (DySPAN07), pp. 168–178, April 2007.

[220] G. Gur and F. Alagoz, “Green wireless communications via cognitive dimension: An overview,” IEEE Network, vol. 25, no. 2, pp.

50–56, March-April 2011.

[221] M. Nekovee, “A survey of cognitive radio access to TV White Spaces,” in Proc. of International Conference on Ultra Modern

Telecommunications & Workshops (ICUMT), Oct. 2009.

[222] C. Ghosh, S. Roy, and D. Cavalcanti, “Coexistence challenges for heterogeneous cognitive wireless networks in TV white spaces,”

IEEE Wireless Communications, vol. 18, no. 4, August 2011, pp. 22-31.

[223] G. P. Villardi, Y. D. Alemseged, C. Sun, C.-S. Sum, T. H. Nguyen, T. Baykas, and H. Harada, “Enabling coexistence of multiple

cognitive networks in TV white space,” IEEE Wireless Communications, vol. 18, no. 4, pp. 32–40, August 2011.

[224] J. Xiao, F. Ye, T. Tian, and R. Q. Hu, “CR enabled TD-LTE within TV white space: System level performance analysis,” in Proc. of

IEEE Global Telecommunications Conference (GLOBECOM), December 2011.

[225] C.-H. Ko, D. H. Huang, and S.-H. Wu, “Cooperative spectrum sensing in TV White Spaces: When cognitive radio meets cloud,” in

Proc. of IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), pp. 672-677, April 2011.

[226] M. A. Rahman, C. Song, and H. Harada, “Development of a TV white space cognitive radio prototype and its spectrum sensing

performance,” in Proc. of International ICST Conference on Cognitive Radio Oriented Wireless Networks and Communications

(CROWNCOM), pp. 231–235, June 2011.

[227] S. Kawade and M. Nekovee, “Can cognitive radio access to TV white spaces support future home networks?,” in Proc. of IEEE

Symposium on New Frontiers in Dynamic Spectrum (DySPAN), April 2010.

[228] A. Zaballos, A. Vallejo, and J. M. Selga, “Heterogeneous communication architecture for the smart grid,” IEEE Network, vol. 25, no.

5, pp. 30–37, September-October 2011.

[229] R. C. Qiu, Z. Hu, Z. Chen, N. Guo, R. Ranganathan, S. Hou, and G. Zheng, “Cognitive radio network for the smart grid: Experimental

system architecture, control algorithms, security, and microgrid testbed,” IEEE Transactions on Smart Grid, vol. 2, no. 4, pp. 724–740,

December 2011.

[230] J. Wang, M. Ghosh, and K. Challapali, “Emerging cognitive radio applications: A survey,” IEEE Communications Magazine, vol. 49,

no. 3, pp. 74–81, March 2011.

[231] A. Ghassemi, S. Bavarian, and L. Lampe, “Cognitive radio for smart grid communications,” in Proc. of IEEE International Conference

on Smart Grid Communications (SmartGridComm), pp. 297–302, October 2010.

[232] R. Yu, Y. Zhang, S. Gjessing, C. Yuen, S. Xie, and M. Guizani, “Cognitive radio based hierarchical communications infrastructure

for smart grid,” IEEE Network, vol. 25, no. 5, pp. 6–14, September-October 2011.

[233] R. Deng, S. Maharjan, X. Cao, J. Chen, Y. Zhang, and S. Gjessing, “Sensing-delay tradeoff for communication in cognitive radio

enabled smart grid,” in Proc. of IEEE International Conference on Smart Grid Communications (SmartGridComm), pp. 155–160,

October 2011.

[234] Q. D. Vo, J.-P, Choi, H. M. Chang, and W. C. Lee, “Green perspective cognitive radio-based M2M communications for smart meters,”

in Proc. of International Conference on Information and Communication Technology Convergence (ICTC), pp. 382–383, November

2010.

[235] C.-K. Yu and K.-C. Chen, “Spectrum map retrieval using cognitive radio network tomography,” in Proc. of IEEE GLOBECOM

Workshops (GC Wkshps), pp. 986–991, December 2011.

[236] K.-D. Lee, S. Kim, and B. Yi, “Throughput comparison of random access methods for M2M service over LTE networks,” in Proc.

of IEEE GLOBECOM Workshops (GC Wkshps), pp. 373–377, December 2011.

Page 64 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

63

[237] H. T. Dinh, C. Lee, D. Niyato, and P. Wang, “A survey of mobile cloud computing: Architecture, applications, and approaches,”

Wireless Communications and Mobile Computing, to appear.

[238] V. Namboodiri, “Towards sustainability in portable computing through cloud computing and cognitive radios,” in Proc. of International

Conference on Parallel Processing Workshops (ICPPW), pp. 468–475, September 2010.

[239] K. Nagothu, B. Kelley, M. Jamshidi, and A. Rajaee, “Persistent Net-AMI for microgrid infrastructure using cognitive radio on cloud

data centers,” IEEE Systems Journal, vol. 6, no. 1, pp. 4–15, March 2012.

[240] C.-H. Ko, D. H. Huang, and S.-H. Wu, “Cooperative spectrum sensing in TV white spaces: When cognitive radio meets cloud,” in

Proc. of IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), pp. 672–677, April 2011.

[241] F. Ge, H. Lin, A. Khajeh, C. J. Chiang, A. M. Eltawil, C. W. Bostian, W.-chun Feng, and R. Chadha, “Cognitive radio rides on the

cloud,” in Proc. of IEEE Military Communications Conference (MILCOM), pp. 1448–1453, October-November 2010.

[242] C. Wang, C. Jiang, S. Tang, and X.-Y. Li, “Scaling laws of cognitive ad hoc networks over general primary network models,” IEEE

Transactions on Parallel and Distributed Systems, vol. 24, no. 5, pp. 1030-1041, May 2013.

[243] O. N. Gharehshiran, A. Attar, and V. Krishnamurthy, “Collaborative sub-channel allocation in cognitive LTE femto-cells: A cooperative

game-theoretic approach,” IEEE Transactions on Communications, vol. 61, no. 1, pp. 325-334, January 2013.

[244] H. ElSawy and E. Hossain, “Two-tier HetNets with cognitive femtocells: Downlink performance modeling and analysis in a multi-

channel environment,” IEEE Transactions on Mobile Computing, to appear.

[245] D. Feng, C. Jiang, G. Lim, L. J. Cimini Jr., G. Feng, and G. Y. Li, “A survey of energy-efficient wireless communications,” IEEE

Communications Surveys & Tutorials, vol. 15, no. 1, pp. 167-178, First Quarter 2013.

[246] S. Huang, H. Chen, Y. Zhang, and H.-H. Chen, “Sensing-energy tradeoff in cognitive radio networks with relays,” IEEE Systems

Journal, vol. 7, no. 1, pp. 68-76, March 2013.

[247] M. Le Treust, S. Lasaulce, Y. Hayel, and G. He, “Green power control in cognitive wireless networks,” IEEE Transactions on Vehicular

Technology, vol. 62, no. 4, pp. 1741-1754, May 2013.

[248] W. Zhong and J. Wang, “Energy-efficient spectrum sharing strategy selection for cognitive MIMO interference channels,” IEEE

Transactions on Signal Processing, vol. 61, no. 14, pp. 3705-3717, July 2013.

[249] T. Han and N. Ansari, “Auction-based energy-spectrum trading in green cognitive cellular networks,” in Proc. of IEEE International

Conference on Communications (ICC 2013), Budapest, Hungary.

[250] M. C. Oto and O. B. Akan, “Energy-efficient packet size optimization for cognitive radio sensor networks,” IEEE Transactions on

Wireless Communications, vol. 11, no. 4, pp. 1544-1553, April 2012.

[251] S. Park, H. Kim, and D. Hong, “Cognitive radio networks with energy harvesting,” IEEE Transactions on Wireless Communications,

vol. 12, no. 3, pp. 1386-1397, March 2013.

[252] F. Granelli, P. Pawelczak, R. Venkatesha Prasad, K. P. Subbalakshmi, R. Chandramouli, J.A. Hoffmeyer, and S. Berger, “Standardization

and research in cognitive and dynamic spectrum access networks: IEEE SCC 41 efforts and open issues,” IEEE Communications

Magazine, January 2010.

[253] “Overview of the IEEE 802.22 Standard on Wireless Regional Area Networks (WRAN) and Core Technologies,” IEEE 802.22 Working

Group Website, http://ieee802.org/22/.

[254] C. Cordeiro, K. Challapali, D. Birru, and N. S. Shankar, “IEEE 802.22: The first worldwide wireless standard based on cognitive

radios,” in Proc. of First IEEE International Symposium on Dynamic Spectrum Access Networks (DySPAN05), pp. 328–337, 8-11

November 2005.

[255] C. Stevenson, G. Chouinard, Z. Lei, W. Hu, S. Shellhammer, and W. Caldwell, “IEEE 802.22: The first cognitive radio wireless

regional area network standard,” IEEE Communications Magazine, vol. 47, no. 1, pp. 130–138, January 2009.

Page 65 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

For Peer Review

64

[256] M. Murroni et al., “IEEE 1900.6 spectrum sensing interfaces and data structures for dynamic spectrum access and other advanced

radio communication systems standard: Technical aspects and future outlook,” IEEE Communications Magazine, vol. 49, no. 12, pp.

118–127, December 2011.

[257] IEEE, “Standard Definitions and concepts for Spectrum Management and Advanced Radio Technologies,” P1900.1 drafts std, v.031,

June 2007.

[258] O. Holland et al., “Development of a Radio Enabler for Reconfiguration Management within the IEEE P1900.4 Working Group,” in

Proc. of IEEE DySPAN 2007, Dublin, Ireland, April 2007.

[259] S. Buljore, V. Merat, H. Harada, S. Filin, and P. Houze, “IEEE P1900.4 system overview on architecture and enablers for optimised radio

and spectrum resource usage,” in Proc. of IEEE symposium on New Frontiers in Dynamic Spectrum Access Networks (DySPAN’08),

October 2008.

[260] SDR Forum, “SDRF-06-P-0009-V1.0.0,” http://www.sdrforum.org/pages/documentLibrary/documents/SDRF-06-P-0009-V1 0 0

CRWG Defs.pdf

[261] ETSI, “ETSI TR 102 838: Reconfigurable radio systems (RRS); Summary of feasibility studies and potential standardization topics,”

http://www.etsi.org, 2009-10.

[262] ITU-R Report 2063, “The Impact of Software Defined Radio on IMT-2000, the Future Development of IMT-2000 and Systems Beyond

IMT-2000.”

[263] ITU-R Report M.2064, “Software-Defined Radio in the Land Mobile Service.”

Page 66 of 66



123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Documents

For Peer Review - Wireless Communications Lab