Doc.: IEEE 802.15-04/022r0 Submission January 2004 Mc Laughlin, decaWave; Welborn, Motorola & Kohno, CRL-UWB Consortium Slide 1 Project: IEEE P802.15 Working

January 2004

Mc Laughlin, decaWave; Welborn, Motorola & Kohno, CRL-UWB ConsortiumSlide 1

doc.: IEEE 802.15-04/022r0

Submission

Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs)Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs)

Submission Title: [Merger#2 Proposal Update ]Date Submitted: [12 January 2004]Source: [Reed Fisher(1), Ryuji Kohno(2), Hiroyo Ogawa(2), Honggang Zhang(2), Kenichi Takizawa(2)] Company [ (1) Oki Industry Co.,Inc.,(2)Communications Research Laboratory (CRL) & CRL-UWB Consortium ]Connector’s Address [(1)2415E. Maddox Rd., Buford, GA 30519,USA, (2)3-4, Hikarino-oka, Yokosuka, 239-0847, Japan] Voice:[(1)+1-770-271-0529, (2)+81-468-47-5101], FAX: [(2)+81-468-47-5431],E-Mail:[(1)[email protected], (2)[email protected], [email protected], [email protected] ]Source: [Michael Mc Laughlin] Company [decaWave, Ltd.]Voice:[+353-1-295-4937], FAX: [-], E-Mail:[[email protected]]Source: [Matt Welborn] Company [Motorola, Inc.]Address [8133 Leesburg Pike, Suite 700, Vienna, Va. 22182, USA]Voice:[+1 703.269.3000], FAX: [+1 703.749.0248], E-Mail:[[email protected]]

Re: [Response to Call for Proposals, document 02/372r8, replaces doc 03/123]

Abstract: []

Purpose: [Summary Presentation of the Merger #2 proposal.]

Notice: This document has been prepared to assist the IEEE P802.15. It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein.Release: The contributor acknowledges and accepts that this contribution becomes the property of IEEE and may be made publicly available by P802.15.

January 2004


doc.: IEEE 802.15-04/022r0

Submission

This Contribution is the Technical Update for Merger #2 Proposal

• This document only contains supplemental information on the Merger #2 proposal. For additional details on the proposal, please see the latest version of the proposal document:

03/334r6 dated November 2003

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Major Contributors For This Proposal Update

Matt WelbornMichael Mc LaughlinJohn McCorkleRyuji KOHNOShinsuke HARAShigenobu SASAKI

Tetsuya YASUIHonggang ZHANGKamya Y. YAZDANDOOSTKenichi TAKIZAWA Yuko RIKUTA

Motorola Inc.decaWave Ltd.Motorola Inc.Yokohama National University Osaka UniversityNiigata University 　

CRL-UWB ConsortiumCRL-UWB ConsortiumCRL-UWB ConsortiumCRL-UWB ConsortiumCRL-UWB Consortium

Supported by:MotorolaMembers of CRL-UWB Consortium

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Presentation Roadmap

• Key differences between DS-UWB and MB-OFDM– Wide occupied bandwidth– Single carrier modulation

• Transmit power calculations

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Big Picture: What are the big differences between DS-UWB & MB-OFDM

• Wideband DS-UWB– Single band occupied by signal (versus frequency hopping)– Compliant with existing regulations– Superior multi-piconet performance– Flexible transmit spectrum

• Single carrier modulation– Excellent performance in indoor multipath channel

• Low fade margins

• Efficient architectures for energy capture

– Scalable to very high data rates

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Signal Occupies Fixed Bandwidth

• Signal continuously occupies widest bandwidth & benefits from UWB advantages– Low fade margin (no Rayleigh fading) in multipath channels– Multipath resolution– Precision ranging

• Other users appear as wideband uncorrelated noise– Offset chip rates for different piconets– Code sequences with low cross-correlation

January 2004


doc.: IEEE 802.15-04/022r0

Submission

3 4 5 6 7 8 9 10 11

High Band

3 4 5 6 7 8 9 10 11

Low Band

3 4 5 6 7 8 9 10 11

Multi-Band

With an appropriate diplexer, the multi-band mode will support full-duplex operation (RX in one band while TX in the other)

Low Band (3.1 to 5.1 GHz) 29 Mbps to 450 Mbps

High Band (6.2 to 10.2 GHz) 29 Mbps to 900 Mbps

Multi-Band (3.1 to 5.1 GHz plus 6.2 GHz to 10.2 GHz) Up to 1.35 Gbps

3 Spectral Modes of Operation

Wide Band DS-UWB

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Compliant With Existing Regulations• In the initial rulemaking, the FCC & NTIA only studied signals

that continuously occupied a single frequency band– Restrictions on gated signals only effective for such signals

– MB-OFDM does not meet this criterion

• APD interference analysis shows that MB-OFMD has identical interference properties to gated UWB signals that are specifically prohibited by the existing rules– Only allowed when power reduced according to duty cycle

• An FCC rule change or interpretation to accommodate MB-OFDM or other FH-UWB waveforms would be needed before certification by FCC

• Deliberations in other regulatory bodies concerned about interference effects of hopped/gated UWB

January 2004


doc.: IEEE 802.15-04/022r0

Submission

APD Analysis for DS-CDMA and MB-OFDMM

.001 0.05 0.1 0.2 0.37-10

-5

0

5

10

15

20

Amplitude Probability Distribution in 50 MHz BW, 250 us Observation

dB

AWGNDS - Root-Raised CosineOFDM3OFDM7OFDM13

The gated OFDM signals have non-Gaussian APDs that indicate large amplitudes with higher probability than for DS-UWB or AWGN

.001 0.01 0.1-10

-5

0

5

10

15

20

Amplitude Probability Distribution in 50 MHz BW, 250 us Observation

dB

11% Gated DSOFDM711% Gated AWGNAWGN

The 11% Gated DS would be specifically prohibited by the UWB rules unless power is reduced by 9.6 dB

The OFDM-7 signal has the same APD and interference properties as the prohibited gated-DS UWB signal

AWGN and noise-like DS-CDMA (Gaussian signals) have flat characteristic curves in an APD plot

Probability of exceeding ordinate0.01 0.05 0.2

Probability of exceeding ordinate

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Superior Multi-piconet Performance• DS-UWB uses FDM and CDM for multi-piconet support

– Low band: 4 full-rate piconets– High band: 4 full-rate piconets (optional)– Both bands: 8 total full-rate piconets (optional)

• Can provide total overlapped SOPs or full duplex operation

• MB-OFDM uses frequency hopping for multi-piconet support– Mode 1: 4 full-rate piconets– Mode 2: 4 full-rate piconets (optional)

• Require use of 3 lowest hop bands, so overlaps Mode I

– Mode 1 + Mode 2: 4 full-rate piconets (optional)• Acquisition occurs in lower 3 bands• Mode 1 and Mode 2 devices operating together provide no

additional SOP benefit (acquisition limited)

January 2004


doc.: IEEE 802.15-04/022r0

Submission

DS-UWB Provides Spectral Flexibility• Signal continuously occupies single wide band

– Data symbols are send serially over entire signal bandwidth

• Receivers are designed to capture energy of transmit pulse through multipath channel– Many receive architectures affected only by difference in Tx power– Receiver performance not affected by Tx pulse shape/spectrum

• As a result, transmit pulse can be modified without any coordination between transmitter and receiver– Flexibility to protect sensitive frequency bands or improve link performance– No resulting changes in data rate, interleaver, etc.– Requires no handshake or message protocol to establish or coordinate

• Provides a path to global harmonization and compliance using optimized UWB pulses

January 2004


doc.: IEEE 802.15-04/022r0

Submission

MB-OFDM Dynamic Bands and Tones Requires Dynamic Coordination

• MB-OFDM proposes that “bands and tones can be dynamically turned on/off” for enhanced coexistence or to meet changing regulations– Dynamically dropping/adding tones or bands would require a message

protocol to dynamically coordinate link parameter changes between transmitter and receiver:

• Dynamic changes in bit-to-carrier tone mapping?• Changes to interleaver? Changes to hopping patterns/codes?• All would require dynamic coordination between transmitters and receivers – No

details have been provided on this mechanism

– Unknown impact on link and piconet performance• Loss of diversity protection against Rayleigh fading for affected bits?• Impact on link performance, data throughput, SOPs, or acquisition?

– MB-OFDM bands with tones “turned off” may not meet the minimum 500 MHz bandwidth (at -10 dB) that is required in the UWB rules

• Prevents or limits “notches” in transmit spectrum

January 2004


doc.: IEEE 802.15-04/022r0

Submission

DS-UWB: Excellent Performance in Indoor Multipath Channel

• Low fade margins– Wide bandwidth results in low fading

– For MB-OFDM, the combination of narrow carriers and punctured FEC degrades performance in multipath channels

• 6 dB or more degradation at 480 Mbps versus AWGN• Cannot be compensated for using equalization or other processing

• Efficient architectures for energy capture– Rake and CMF architectures provide efficient and scalable energy

capture

• Equalization compensates for inter-symbol interference– Proven & widely used equalizer technology

January 2004


doc.: IEEE 802.15-04/022r0

Submission

DS-UWB in Multipath

• Indoor multipath channels provide several challenges for UWB systems– Multipath fading– Inter-symbol interference (ISI)– Energy capture

• Effects are well-understood and are analyzed as trade-off between performance versus complexity

• DS-UWB minimizes fading and provides scalable energy capture

• MB-OFDM provides good energy capture at the expense of significant multipath fading

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Compensating for ISI• ISI occurs as a result of non-uniform channel frequency response

– Multipath delay spread exceeds the symbol interval

• ISI is compensated for using an equalizer– Linear equalizer (digital filter)– Decision-feedback equalizer (DFE)

• If left uncompensated, ISI can cause high BER & error floor phenomenon

• Equalizer technology is widely used in many types of systems– Telephone modems– WLAN (e.g. 802.11b)– HDTV

• OFDM systems use frequency domain equalization to compensate for phase and amplitude response of channel– If delay spread exceeds CP length, residual ISI compensation would

require additional time-domain equalization

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Example of ISI Effects on BER

5 10 15 20-6

-5

-4

-3

-2

-1

Eb/No

Lo

g10

BE

RUncoded Equalization Performance on CM3-15, 16 Finger Rake

No Equalization9-Tap Least-Squares DFEAWGN Channel

• Un-equalized system experiences high BER and error floor• 9-tap DFE with 16-finger rake performance approaches AWGN

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Equalizer Design Trade-offs• Equalizer is a key part of any modern wireless receiver architecture

– Phase correction and MRC in an MB-OFDM receiver are equivalent to equalization in a DS-UWB receiver – without equalization, MB-OFDM receiver would fail to operate

– 8-bit channel estimate required for MB-OFDM

• System design and simulation for DS-UWB without an equalizer results in sub-optimal performance

– Degraded SNR due to uncompensated ISI– Presence of error floors due to uncompensated ISI– Poor narrow-band interference rejection and multi-piconet performance due

to degraded SNR

• Complexity analysis involves many assumptions for a particular implementation – equalization is only one of them

• Complexity analysis without an equalizer results in excessive implementation complexity

– Example: sub-optimal performance without an equalizer results in excessive bit-widths or rake taps and therefore unnecessary complexity

– Small number of gates for equalizer may result in big complexity reduction

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Eye diagram at Eb/No = 5 dB Noise dominates ISI

5 10 15 20-6

-5.5

-5

-4.5

-4

-3.5

-3

-2.5

-2

-1.5

-1

Eb/No

Log1

0 B

ER

Uncoded Equalization Performance on CM3-15, 16 Finger Rake


100 200 300 400 500 600 700 800 900 1000 1100

-1

-0.5

0

0.5

1

16 Rake, No DFE, Eb/No = 5

100 200 300 400 500 600 700 800 900 1000 1100

-1

-0.5

0

0.5

1

16 Rake, Post DFE, Eb/No = 5

Received BPSK Symbols

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Eye diagram at Eb/No = 10 dBNoise/ISI at similar levels

100 200 300 400 500 600 700 800 900 1000 1100

-1

-0.5

0

0.5

1


100 200 300 400 500 600 700 800 900 1000 1100

-1

-0.5

0

0.5

1



5 10 15 20-6

-5.5

-5

-4.5

-4

-3.5

-3

-2.5

-2

-1.5

-1

Eb/No

Log1

0 B

ER



January 2004


doc.: IEEE 802.15-04/022r0

Submission

Eye diagram at Eb/No = 15 dB ISI dominates AWGN

5 10 15 20-6

-5.5

-5

-4.5

-4

-3.5

-3

-2.5

-2

-1.5

-1

Eb/No

Log1

0 B

ER



100 200 300 400 500 600 700 800 900 1000 1100

-1

-0.5

0

0.5

1


100 200 300 400 500 600 700 800 900 1000 1100

-1

-0.5

0

0.5

1



January 2004


doc.: IEEE 802.15-04/022r0

Submission

Eye diagram at Eb/No = 20 dB ISI dominates AWGN

100 200 300 400 500 600 700 800 900 1000 1100

-1

-0.5

0

0.5

1


100 200 300 400 500 600 700 800 900 1000 1100

-1

-0.5

0

0.5

1



5 10 15 20-6

-5.5

-5

-4.5

-4

-3.5

-3

-2.5

-2

-1.5

-1

Eb/No

Log1

0 B

ER



January 2004


doc.: IEEE 802.15-04/022r0

Submission

Simulations without EqualizersWithout Equalizer With Equalizer

Uncompensated ISI leads to degraded SNR

Equalizer improves output SNR longer range & lower BER

Uncompensated ISI leads to “error floor” in simulation

Equalizer removes error floor

Uncompensated ISI leads to degraded NBI rejection from degraded SNR

Equalizer improves NBI rejection performance

Poor error performance without equalizer leads to excessive complexity estimates

Example: Design for 4-bit rake taps to achieve desired performance

Improved performance allows lower complexity implementation

Example: fewer bits needed for taps

800 gate multiplier 400 gates

204K gate rake 102K gate rake

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Multipath Performance Conclusions• Energy capture and ISI compensation are SOLVED PROBLEMS for

DS-UWB systems– Both are simple a complexity versus performance trade-off– Energy capture degradation is ~1 dB (or less) for 16-finger rake in CM1-3

and less CMF for modest complexity (regardless of data rate)– 64-BOK modes outperform MB-OFDM without an equalizer, performance

would improve with appropriate equalization

• Performance scales with process technology– CMF, rake & equalizer require fewer gates and/or less power in faster &

smaller process

• Rayleigh fading for MB-OFDM cannot be mitigated by any amount of added signal processing– High rate modes degraded by 6 dB or more relative to AWGN– Rayleigh fading performance does not improve with process technology or

added digital processing

January 2004


doc.: IEEE 802.15-04/022r0

Submission

PHY Comparisons • Previous example shows pitfalls of comparing proposals based only on

simulations of a few specific implementations– Specific implementation decisions can lead to significant differences in performance

and complexity estimates– Sub-optimal implementations lead to poor/irrelevant simulation results– Technology development and design optimization will lead to improved

implementations

• Best decision will also be based on fundamental or asymptotic performance bounds seasoned with experience and judgment– Leads to a standard that can continue to provide improved performance-to-cost as

technology matures

• DS-UWB provides scalable performance that exceeds MB-OFDM today and continues to improve with more sophisticated implementations– Scalable energy capture, ADC bit widths & equalizer performance– Scales to higher data rates without Rayleigh fading losses – Multi-user detection possible for improved SOP operation

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Transmit Power Calculations

• UWB average transmit power is constrained by FCC regulations based on power spectral density– Average power limit equivalent to -41.25 dBm/MHz – Peaks in spectrum above “average” require reduction in transmit power

to meet limit at highest power frequency

• MB-OFDM proposed transmit power is -10.3 dBm– -41.25 dBm+10*Log(4.125 MHz*300 data carriers) dB = -10.3 dBm– Assumes perfectly flat spectrum for each of the 300 QPSK-modulated

carriers (100 data carriers each in 3 bands) at FCC limit– Used as basis for performance in all link budgets and simulations

• All simulations to date appear to assume 0 dB transmit power back-off

• Any non-zero ripple in transmit spectrum would require reduction in transmit power– Would result in shorter ranges or worse BER performance

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Realistic Measurement of Transmit Power• Claim of perfectly flat MB-OFDM spectrum is based on theoretical result

for zero-padded prefix OFDM over infinite signal duration

• Real limits will be based on spectrum analyzer measurements of transmit power during FCC certification testing

– 1 MHz resolution bandwidth, RMS averaging, peak-hold

• Real spectrum analyzer measurement with 1 MHz RBW and RMS averaging results in ripple for “theoretically flat” MB-OFDM signals

– 1 MHz RBW effectively integrates only about 1 microsecond of signal– For MB-OFDM, this is little more than 1 symbol of data– Carriers in each symbol are just sinusoids with “random” phases – not QPSK – RMS averaging discards phase information so averaged carriers are not

spread by multiple symbols of modulation with different phases– Instead, carriers begin to be “resolved” by 1 MHz RBW and RMS averaging

• DS-UWB signal has > 1000 pulses in 1 micro-second observation– Spectrum in 1 MHz RBW is well-spread BPSK/QPSK with shape determined

only by transmit pulse and MBOK code sequence

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Realistic Measurement of Transmit Power• Calculations of MB-OFDM spectrum were made that show ~1 db

of ripple in MB-OFDM power spectrum• Transmit power will need to be reduced by about 0.5 dB below

the proposed level of -10.3 dBm to compensate for spectral ripple

• FCC testing based on current policy for gated UWB signals (equivalent to MB-OFDM interference characteristics) also requires power reduction based on signal duty cycle (e.g. 25% duty cycle 6 dB attenuation)

• No definitive interpretation has yet been issued by FCC WRT measurement procedures for frequency-hopped UWB

Spectrum Analysis

1 MHz RBWFilter

RMS Detector

MB-OFDM SignalGeneration

Plot of PowerSpectrum

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Radiated Emissions UWB

Radio Sample 2

Test Distance: 1mDetector: RMSRBW/VBW: 1 MHz/3 MHzMeas. Time: 1 msEmissions: < Limit

Note: Data normalized to 3m test environment for limit comparison.

Apparent ripple of 6 dB or more

*Plot from MBOA-reported spectrum measurements (assumed to have CP)

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Calculated MB-OFDM Spectrum

Calculation Parameters

Zero-padded Cyclic PrefixDetector: RMSRBW: 1 MHz

Spectrum showing non-zero ripple

*Spectrum plot based on analysis using FCC measurement rules

MB-OFDM Band 2 Signal Averaged over 1 ms

Frequency (MHz)

Nor

mal

ized

Pow

er S

pect

rum

(dB

)

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Transmit Power Calculations• DS-UWB power based on RRC pulse shape + stated

transmit power back-off to account for code sequence– -41.25 dBm + 10*Log(1368 MHz) = -9.9 dBm – 1368 MHz is 3 dB bandwidth (RRC spectrum is symmetric

about 3 dB point on a linear scale)

• Original transmit back-off numbers based on code sequence analysis (included 16-BOK requirement)

• Updated analysis results in lower back-off numbers:– 2-BOK requires worst-case 1.9 dB back-off– 4-BOK requires worst-case 1.2 dB back-off– 8-BOK requires worst-case <1 dB back-off– 64-BOK requires worst-case 0.4 dB back-off

January 2004


doc.: IEEE 802.15-04/022r0

Submission

MB-OFDM Power Assuming Favorable UWB Rules Change or Interpretation

DS-UWB

4-BOK

DS-UWB

64-BOK

MB-OFDM Mode I

Theoretical transmit power -9.9 dBm -9.9 dBm -10.3 dBm

Transmit back-off 1.2 dB 0.4 dB 0.5 dB

Attenuation for FCC compliance 0 dB 0 dB 0 dB*

Actual transmit power -11.1 dB -10.3 dB -10.8 dB

Path loss 44.4 dB 44.4 dB 44.2 dB

Received power -55.5 dBm -54.7 dBm -55.0 dBm

Received power WRT MB-OFDM -0.5 dB +0.3 dB 0 dB

*Assumes best-case favorable UWB rules change or interpretation to allow higher power frequency-hopped UWB despite gated-UWB interference characteristics

January 2004


doc.: IEEE 802.15-04/022r0

Submission

MB-OFDM Power Assuming Limits Based on Gated-UWB Interference Characteristics

DS-UWB

4-BOK

DS-UWB

64-BOK

MB-OFDM Mode I

Theoretical transmit power -9.9 dBm -9.9 dBm -10.3 dBm

Transmit back-off 1.2 dB 0.4 dB 0.5 dB

Attenuation for FCC compliance 0 dB 0 dB -5.9 dB*

Actual transmit power -11.1 dB -10.3 dB -16.7 dB

Path loss 44.4 dB 44.4 dB 44.2 dB

Received power -55.5 dBm -54.7 dBm -60.9 dBm

Received power WRT MB-OFDM +5.4 dB +6.2dB 0 dB

*Assumes current FCC measurement procedure for frequency-hopped UWB based on gated-UWB power attenuation requirements (attenuation based on 26% duty cycle for Mode I – Mode II higher)

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Consequences of Revised Transmit Back-off Results

• DS-UWB transmit back-off analysis leads to higher transmit power– 0.6 dB higher for 64-BOK 7% range increase (if noise limited)

– 0.9 dB higher for 4-BOK 11% range increase (if noise limited)

• Calculated MB-OFDM spectral ripple leads to lower transmit power and decreased range– 0.6 dB lower Tx power 7% range reduction

• Relative gain with Tx power back-off adjustments:– 4-BOK vs. MB-OFDM: +1.4 dB

– 64-BOK vs. MB-OFDM: +1.1 dB

• MB-OFDM power reduction to account for FCC compliance without favorable rules change/interpretation leads to 5.9 dB reduction– Total 6.4 dB power reduction 52% range reduction

January 2004


doc.: IEEE 802.15-04/022r0

Submission

Conclusions• Primary differences between DS-UWB and MB-OFDM

– Use of a single continuously occupied band

– Use of single carrier modulation

• ISI compensation through equalization is well understood and widely used– Can lead to significant performance improvements for DS-UWB

• PHY evaluation should also emphasize fundamental limitations– Example: limited scalability for MB-OFDM to higher data rates due

to Rayleigh fading in multipath channels

• Realistic transmit power calculations will lead to more reliable range and performance predictions

Documents

Doc.: IEEE 802.15-04/022r0 Submission January 2004 Mc Laughlin, decaWave; Welborn, Motorola & Kohno, CRL-UWB Consortium Slide 1 Project: IEEE P802.15 Working