Embed Size (px)
Citation preview
Cape Town, South Africa 12–14 December 2011
Dhananjay Kumar, M.E., M.Tech., Ph.D. Anna University, Chennai [email protected]
Kanagaraj N. N., B.Tech.
Alcatel-Lucent India Limited, Chennai
Radio Resource Management in OFDMA-CRN Considering Primary
User Activity and Detection Scenario
ITU Kaleidoscope 2011 The fully networked human?
Innovations for future networks and services
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Outline
Cognitive Radio Environment
Real-time Observations & Requirement Analysis
Learning in Cognitive Radio Networks
Resource Allocation in Multi-carrier Systems
Problem Formulation in OFDMA-CRN
Simulation Results
Conclusion & Future Works
Cognitive Radio Environment
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
BASE STATION
CR USER 1
Cognitive Network (Secondary)
CR USER 3
CR USER 1
CR USER 4
CR USER 2
MESH ROUTER 1 MESH ROUTER 3
MESH ROUTER 2
PRIMARY USER 2
PRIMARY USER 3
PRIMARY USER 1
TV Broadcast Network
(Primary)
PRIMARY USER 4
PRIMARY USER 5
IEEE 802.11n Access Point (ISM Band)
Wireless LAN (Primary)
Wireless Mess Network (Primary)
Cognitive Link
Primary Link
Opportunity in 900 MHz Band
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Measurements taken at Adyar, Chennai on 16th August, 2011
Fig.2. At 12:24:44 p.m.
Fig.1. At 12:14:12 p.m.
Fig.3. At 12:49:13 p.m.
Spectrum Usage in 2.4 GHz Band
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Fig.1. At 11:30:43 a.m.
Fig.2. At 11:39:31 a.m.
Fig.3. At 11:39:51 a.m.
Fig.4. At 11:40:39 a.m.
On 24th August, 2011 (Adyar, Chennai)
Learning in Cognitive Radio Network
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Learning Algorithms
PARAMETERS RULES ENGINE
KNOWLEDGE
WAVEFORM INTERNAL STATE
EXTERNAL STATE
CONTROL
OPERATIONAL FUNCTIONS
SENSORS
RADIO SYSTEM
TX TX RX
WIRELESS ENVIRONMENT
RADIO COMMUNICATION CHANNELS
Learning in a Broader Perspectives
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Ref: Joseph Gaeddert et al., “Applying Artificial Intelligence to the Development of a Cognitive Radio Engine”, http://wireless.vt.edu/archives/download/ApplyingArtificialIntelligence.pdf
DOMAIN KNOWLEDGE BASE
CASE MEMORY
RADIO ENVIRONMENT
MAP(REM) OBSERVATION
HMM, DATA MINING
GENETIC ALGORITHMS
SEARCH BASED
LEARNING
NEURAL NETWORK
COOPERATIVE LESRNING
Prediction, planning
Parameter optimization,
tradeoff
Collaboration with other nodes
Learning and Reasoning
Situation Awareness and
Information Base
Decision and Adaptation
KNOWLEDGE BASED
LEARNING
CASE BASED
LEARNING
Resource Allocation in Multi-carrier Base Cognitive Radio Networks
Is situation-aware learning is enough in multi-carrier systems?
Resource allocation
Sub-carrier selection
Power allocation
Resource optimization
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Resource Allocation in OFDMA-CRN: Problem Formulation
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Objective Function:
Optimization Methods
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Optimization Methods cont..
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Sub-carrier and Power Allocation
(SPA) Algorithm
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Start
Categorize the secondary users into different groups
Initialize (λ1, λ2, . . .λN)
k=1(subcarrier)
Compute P*g, k for all groups
Pick the group g* that gives the minimum of Dk(λ)
Assign the sub-carrier to the group g*
Allocate power for the group g*
Set Pg*, k = P*g*, k and Pg, k =0 for all g ≠ g*
Increment k
Stop
Is k<K Yes
No
Main Simulation Parameters
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Parameters Description Value
N Number of primary users 2
K Number of OFDM sub-carriers 128
B Maximum spectrum hole 10MHz
G Number of groups 10
|Mg| Number of secondary users in
each group g
4
Effects of Varying Loss Parameters
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10.4
0.41
0.42
0.43
0.44
0.45
0.46
0.47
0.48
0.49
Phi (with fixed beta = 0.03)
Data
rate
(bps p
er
Hert
z)
C1 = 0.0
C1 = 0.2
C1 = 0.4
C1 = 0.6
C1 = 0.8
C1 = 1.0
Effects of Varying Detection Parameters
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10.44
0.46
0.48
0.5
0.52
0.54
0.56
0.58
0.6
Beta (with fixed phi = 0.04)
Data
rate
(bps p
er
Hert
z)
C2 = 0.0
C2 = 0.2
C2 = 0.4
C2 = 0.6
C2 = 0.8
C2 = 1.0
Sum Data Rates in Two Proposed Scenarios
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
2 3 4 5 6 7 8 9 100.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
No. of groups (secondary users)
Data
rate
(bps p
er
Hert
z)
Scenario 1(phi=0.08,beta=0.05)
Scenario 2(phi=0.08,beta=0.05)
Scenario 1(phi=0.04,beta=0.03)
Scenario 1(phi=0.04,beta=0.03)
Throughput Comparison
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
2 3 4 5 6 7 8 9 1016
17
18
19
20
21
22
23
24
No. of groups (secondary users)
Thro
ughput(
Mbps)
Scenario 1 (phi=0.08,beta=0.05)
Scenario 2 (phi=0.08,beta=0.05)
Scenario 1 (phi=0.04,beta=0.03)
Scenario 2 (phi=0.04,beta=0.03)
Assessment of Processing Delay
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
2 3 4 5 6 7 8 9 100.006
0.008
0.01
0.012
0.014
0.016
0.018
0.02
0.022
No. of groups (secondary users)
Pro
cessin
g d
ela
y(s
econds)
Scenario 1
Scenario 2
Conclusion & Future Work
Analysed the spectrum availability based on real-time measurements.
Discussed the learning scenarios
Defined the objective function and optimized it analytically.
Developed Sub-carrier and Power Allocation (SPA) Algorithm for OFDMA-CRN.
Analysis & simulation of the effect of both issues i.e. primary user activity and detection
Unification of learning algorithm with sub-carrier allocation in OFDMA-CRN is our future work.
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
Thank You
Cape Town, South Africa, 12-14 December 2011 ITU Kaleidoscope 2011 – The fully networked human? Innovations for future networks and services
