1. WWiSE IEEE 802.11n Proposal Authors: Date: 2005-01-06
Notice: This document has been prepared to assist IEEE 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 grants a free, irrevocable license to the IEEE to incorporate material contained in this contribution, and any modifications thereof, in the creation of an IEEE Standards publication; to copyright in the IEEE’s name any IEEE Standards publication even though it may include portions of this contribution; and at the IEEE’s sole discretion to permit others to reproduce in whole or in part the resulting IEEE Standards publication. The contributor also acknowledges and accepts that this contribution may be made public by IEEE
Patent Policy and Procedures: The contributor is familiar with the IEEE 802 Patent Policy and Procedures < ieee802.org/guides/bylaws/sb-bylaws.pdf>, including the statement "IEEE standards may include the known use of patent(s), including patent applications, provided the IEEE receives assurance from the patent holder or applicant with respect to patents essential for compliance with both mandatory and optional portions of the standard." Early disclosure to the Working Group of patent information that might be relevant to the standard is essential to reduce the possibility for delays in the development process and increase the likelihood that the draft publication will be approved for publication. Please notify the Chair as early as possible, in written or electronic form, if patented technology (or technology under patent application) might be incorporated into a draft standard being developed within the IEEE Working Group. If you have questions, contact the IEEE Patent Committee Administrator at
Sean Coffey, Texas Instruments, et al.

2. Authors, contd:
Sean Coffey, Texas Instruments, et al.

3. Abstract The presentation summarizes the WWiSE complete proposal:
Recap of proposal features
Updates since last meeting
Performance and comparisons with other proposals
New technical supporting data on several aspects of the WWiSE design
Sean Coffey, Texas Instruments, et al.

4. The WWiSE approach WWiSE = World Wide Spectrum Efficiency
The partnership was formed to develop a specification for next generation WLAN technology suitable for worldwide deployment
Mandatory modes of the WWiSE proposal comply with current requirements in all major regulatory domains: Europe, Asia, Americas
Proposal design emphasizes compatibility with existing installed base, building on experience with interoperability in g and previous amendments
All modes are compatible with QoS and e
Maximal spectral efficiency translates to highest performance and throughput in all modes
Sean Coffey, Texas Instruments, et al.

5. Summary of proposal features
WWiSE proposes 2 transmitters in 20 MHz mandatory
Rates 54, 81, 108, 121.5, 135 Mbps
Optional 40 MHz counterparts of all 20 MHz modes
Optional extensions to 3 and 4 transmit antennas
Optional space-time block codes for longer range
Optional LDPC code
MAC: HTP burst, aggregation, extended Block Ack
See r3 & r4 for a full description
Sean Coffey, Texas Instruments, et al.

6. Unified format 1/2 800 BCC, LDPC 16-QAM 3/4 2/3 64-QAM 5/6 Code rate
Cyclic prefix, ns
Code
Constellation
1/2
800
BCC, LDPC
16-QAM
3/4
2/3
64-QAM
5/6
All combinations of 2, 3, 4 transmit antennas and 20/40 MHz offer exactly these 5 modes
All 20 MHz modes have 54 data subcarriers, 2 pilots. All 40 MHz modes have 108 data subcarriers, 4 pilots
Sean Coffey, Texas Instruments, et al.

7. 2 transmitter SDM, 20 MHz (mandatory)
PHY rate
Data carriers
Pilots
Code rate
Cyclic prefix, ns
Code
Constellation
54 Mbps 54 2
1/2
800
BCC
16-QAM
81 Mbps
3/4
108 Mbps
2/3
64-QAM
121.5 Mbps
135 Mbps
5/6
Sean Coffey, Texas Instruments, et al.

8. Optional data modes 20 MHz: 3 Tx space-division multiplexing
40 MHz: (all 40 MHz modes optional)
1 Tx antenna
2 Tx space division multiplexing
3 Tx space division multiplexing
Space-time block codes: (all STBCs optional)
2x1, 3x2, 4x2, 4x3 in 20 MHz and 40 MHz
LDPC code option
Optional in all MIMO configurations and channel bandwidths
Sean Coffey, Texas Instruments, et al.

9. Optional mode data rates, multiple spatial streams
20 MHz:
Configuration
Rate ½, 16-QAM
Rate ¾, 16-QAM
Rate 2/3, 64-QAM
Rate ¾, 64-QAM
Rate 5/6, 64-QAM
3 Tx, 20 MHz 81
121.5
162
182.25
202.5
4 Tx, 20 MHz
108
216
243
270
40 MHz:
Configuration
Rate ½, 16-QAM
Rate ¾, 16-QAM
Rate 2/3, 64-QAM
Rate ¾, 64-QAM
Rate 5/6, 64-QAM
2 Tx, 40 MHz
108
162
216
243
270
3 Tx, 40 MHz
324
364.5
405
4 Tx, 40 MHz
364
432
486
540
Sean Coffey, Texas Instruments, et al.

10. Optional mode data rates, single spatial stream
2x1, 20 MHz:
1x1, 40 MHz; 2x1, 40 MHz
PHY rate, Mbps
Code rate
Constellation
6.75
1/2
BPSK
10.125
3/4
13.5
QPSK
20.25 27
16-QAM
40.5 54
2/3
64-QAM
60.75
67.5
5/6
PHY rate, Mbps
Code rate
Constellation
13.5
1/2
BPSK
20.25
3/4 27
QPSK 41 54
16-QAM 81
108
2/3
64-QAM
121.5
135
5/6
Sean Coffey, Texas Instruments, et al.

11. Update Information exchange mechanism added Rate & mode recommendation
Channel state information
10 MHz channelization supported
All 20 MHz modes have a ½-data rate 10 MHz counterpart
ZIFS (zero interframe space) mode deleted
HTP Burst already allows this in effect
Sean Coffey, Texas Instruments, et al.

12. Update, contd. Rate & mode recommendation
It is of critical importance that this information is advisory and does not mandate Tx behavior
Rate selection algorithms do not need to be redesigned
There is no need for an elaborate protocol to decide when information is stale
The transmitter (e.g., AP) may in many situations have more information about overall network conditions than the receiver, should be able to override receiver request
Facilitates low power operation
E.g., in receiver that is at the edge of its capabilities at the higher data rate
Sean Coffey, Texas Instruments, et al.

13. Update, contd. Channel state information exchange:
General purpose mechanism, built on already existing mechanisms in h
Sufficient precision for current and future purposes
Sean Coffey, Texas Instruments, et al.

14. Comparisons and contrasts
Sean Coffey, Texas Instruments, et al.

15. The context within the task group
Common points across the complete proposals:
2 transmitter MIMO-OFDM
20 MHz
Open loop
Varying evolutions to the ag OFDM format
Data rates significantly in excess of 2x54 Mbps
Aggregation of varying kinds
Block acknowledgements
Sean Coffey, Texas Instruments, et al.

16. Context within the task group, contd.
Technical data:
Extensive new data has been brought by many parties
While there are clearly areas of disagreement, there are several ways in which there is some emerging agreement on the data
This presentation will attempt to distinguish between areas of at least some agreement, and areas still outstanding
Sean Coffey, Texas Instruments, et al.

17. “Scalable Architecture across several dimensions” (adapted from 11-04-0888r3/slide 5 [TGnSync])
Market segments
“The same HT technology able to function in all markets and environments”
- Wi-Fi Alliance MRD, Nov.’03 PC
Enterprise
Consumer Electronics
Public Access
Handset
WWiSE rates; applicable in all listed market segments
135Mbps
270Mbps
540Mbps
Perf.
over
time
140Mbps
315Mbps
630Mbps
10MHz (.11j/p)
20MHz
In WWiSE; unified format means exact match of offerings across bandwidths
Regulatory Domains
40MHz
Sean Coffey, Texas Instruments, et al.

18. “… And a well-defined Core” (adapted from 11-04-0888r3/slide 6 [TGnSync])
Market segments
Mandatory WWiSE features:
Two antennas
20 MHz
135 Mbps
Mandatory Features:
Two antennas
20MHz
140 Mbps
(TGnSync)
Perf .
over
time
Regulatory Domains
Sean Coffey, Texas Instruments, et al.

19. Design principles – I Range and robustness:
End-user expectations of range and coverage are driven by experience with existing ag equipment
It is critical that modes offered in 11n have ranges comparable to or matching the ranges obtained by 54 Mbps ag
Sean Coffey, Texas Instruments, et al.

20. Design principles – II One HT technology:
“The same HT technology should work equally well in all market segments [residential, enterprise, hotspot].”
Wi-Fi Alliance MRD, November 2003, r2
HT technology should work equally well on a worldwide basis
Sean Coffey, Texas Instruments, et al.

21. Design principles – III
Open-loop core operation:
is successful as a flexible, distributed, open-loop standard
Extensions build traction on the basis of the core, rather than redefining the basic model or impairing its performance
Closed loop operation, like polling, may become a successful and widespread extension
Sean Coffey, Texas Instruments, et al.

22. Technical issues Issues affecting range, coverage and robustness:
Optional advanced coding
Maximum FEC code rate
Useful bandwidth in 20 MHz
Pilot tone usage
Cyclic prefix
Interleaver design
Preamble design:
Legacy compatibility
Beamforming and preamble design
Sean Coffey, Texas Instruments, et al.

23. Range issues – optional advanced coding
Both WWiSE & TGnSync propose optional LDPC codes
Not the same code design
Gains from LDPC codes can be expected to be 2 dB to 2.5 dB
Note that this comes at an appreciable cost, certainly in the several 100,000’s of gates
Detailed comparisons of performance and complexity are pending
Sean Coffey, Texas Instruments, et al.

24. Range issues – FEC code rate limit
Increasing the FEC code rate always affects range
Unlike all the other (open-loop) ag format changes under consideration
Designs have increased FEC code rates to the point of criticality
Robustness depends importantly on redundancy
Increasing from code rate 5/6 to 7/8 reduces fraction of redundant bits 25%, while only increasing data rate 5%
When combined with 64-QAM & MMSE detection, this change costs 2.1 dB in capacity
Increasing from code rate 3/4 to 7/8 costs over 5 dB
Sean Coffey, Texas Instruments, et al.

25. Range issues – useful 20 MHz bandwidth
WWiSE proposes increasing number of tones used from 52 to 56
Fits the existing spectral mask; see r3 and -r4.
Negligible effect on ACI; see new technical information in r0, “Adjacent channel interference and filtering for 56-carrier signals,” Dave Hedberg.
Direct effect on range and robustness, through code rate required to meet given data rate
For example, 108 Mbps may be achieved either by rate 3/4, 52 tones (48 data, 4 pilots), or by rate 2/3, 56 tones (54 data, 2 pilots)
Approximately 2 dB gain
Sean Coffey, Texas Instruments, et al.

26. Range issues – pilot tone usage
WWiSE proposes using 2 pilots instead of 4
2 pilots into 2 Rx antennas provide similar or better tracking performance to 4 pilots into 1 antenna, as in ag
Simulations ( r0, Allert van Zelst et al.) confirm that overall system performs excellently: approx. 0.5 dB degradation relative to ideal tracking (infinitely many pilots)
Efficient pilot usage eases strain on FEC code rate; cf. 4% pilot improvement & 5% difference between rates 5/6 and 7/8
A pilot test: in any system without reduced pilots, compare performance with a revised system in which code rate is reduced to match efficient pilots
Sean Coffey, Texas Instruments, et al.

27. Range issues – pilot tone usage, contd.
Other simulations ( r0, Won-Joon Choi et al.):
Also show low degradation when there are two receivers;
Expressed concern when there is one spatial stream, or narrowband interference
Comments:
Any two-receive-antenna system will have excellent tracking, even with one spatial stream
The simulations in question did not implement averaging over consecutive symbols; see for resulting improvement
The narrowband interference simulated is unrealistic; and in any case should be compared with performance loss when data tones are hit equivalently (cf. the “pilot test”, previous slide)
Sean Coffey, Texas Instruments, et al.

28. Range issues – shortened cyclic prefix
TGnSync propose halving cyclic prefix to 400 ns, as a mandatory mode
Cyclic prefix must be long enough to absorb multipath in channel, and spread caused by Tx and Rx filtering
TGnSync appear to apply this only to channel model B;
Essentially no results are presented for channel models D & E
No modifications to filtering are made to accommodate this mode
Per r0, “TGnSync response to Monterey questions”
Sean Coffey, Texas Instruments, et al.

29. “Scalable Architecture across several dimensions” (adapted from 11-04-0888r3/slide 5 [TGnSync])
Market segments
Each of these TGnSync rates has short cyclic prefix and code rate 7/8
- apparently will not work at all in
enterprise & public access
- will not be robust in any segment PC
Enterprise
Consumer Electronics
Public Access
Handset
Perf.
over
time
140Mbps
315Mbps
630Mbps
10MHz (.11j/p)
20MHz
Regulatory Domains
40MHz
Sean Coffey, Texas Instruments, et al.

30. Range issues – MIMO interleaver
The WWiSE interleaver outperforms the original TGnSync interleaver substantially
Approximately 1.25 dB gain, with no offsetting advantage for TGnSync design; see r4, slides 19-20
Subsequent to the last IEEE meeting, TGnSync have changed their interleaver
The new structure follows the basic setup of the WWiSE interleaver; the parameters used are slightly different
This reduces the gap in performance, but does not quite eliminate it, per initial simulations; further work pending
Sean Coffey, Texas Instruments, et al.

31. MIMO interleaving (Cf. 11-04-0935r3, slide 29 [WWiSE])
TGnSync change subcarrier shift and Idepth parameter
Bit-cycling across NTX transmitters
Parameterized a-style interleaver
5 subcarrier shift, same interleaver .
TX 0 interleaved bits
Coded bits
TX 1 interleaved bits
Configuration
Idepth
108 tones, 1 Tx, 2x1 12
All others 6
Shift of 5 additional subcarriers for each additional antenna
Sean Coffey, Texas Instruments, et al.

32. Preamble design – legacy compatibility
Experimental results on effect of WWiSE short sequence on several devices show excellent performance
See r0, Tushar Moorti et al.
Equipment tested came from Atheros, Broadcom, Conexant, and TI
This presentation also outlines flexibility in the design
The chosen short sequence cyclic shifts were best in terms of power step, but several others work well also
Sean Coffey, Texas Instruments, et al.

33. Preamble design – legacy compatibility, contd.
Doc r6, slides 33-38, describes results from another vendor, indicating loss of performance
We believe that there should be no fundamental barrier to a short sequence of the general proposed form working
Further experimental results are pending.
Sean Coffey, Texas Instruments, et al.

34. Preamble design – compatibility with beamforming
The WWiSE proposal achieves an efficient, compact long sequence design
This uses regularity existing in realistic channels to extract channel estimates
This is a proven technique for improving channel estimates that is well documented in open-source literature
Beamforming operates on the same channels and retains much of the same regularity
Simulation results in r0, Chris Hansen et al., demonstrate use of the WWiSE preamble with beamforming
Excellent channel estimation results
Sean Coffey, Texas Instruments, et al.

35. Other comparisons Simplicity, scalability, modularity:
In the WWiSE proposal, any modes offered in 20 MHz are offered unchanged at 10 MHz & 40MHz;
Any combination of features offered in one antenna configuration is offered in all
Range-enhancing space-time block codes fit in a modular way onto the basis formed by the rest of the proposal
Any closed loop methods would fit in exactly the same way
Sean Coffey, Texas Instruments, et al.

36. Other comparisons, contd.
Optionality of 40 MHz:
The standard uses options when there is some reason why it is inappropriate to insist on mandatory behavior
In the case of n, there are already optional modes in common across the proposals that fit this criterion:
3 and 4 transmit antennas
Advanced coding (LDPC codes)
In these cases, the issue preventing immediate adoption is cost
In the case of 40 MHz channels, these are prohibited in major regulatory domains and should also be optional
Sean Coffey, Texas Instruments, et al.

37. Other comparisons, contd.
Optionality of 40 MHz, contd.:
40 MHz channels are currently prohibited in Japan
Europe has very many different regulatory domains
40 MHz channels do not conform to ETSI format
Applications develop around standards; fundamentally different flavors will fragment the positive benefits that are the point of the standardization procedure
There is no technical barrier in the WWiSE proposal to making 40 MHz mandatory
We simply believe (unanimously) that mandatory 40 MHz is very bad policy
Sean Coffey, Texas Instruments, et al.

38. New technical information - I
Adjacent channel interference:
r0, “Adjacent channel interference and filtering for 56-carrier signals,” Dave Hedberg.
Requirements for Tx and Rx filtering for the 56 carrier WWiSE signals do not increase complexity significantly e.g. anti-aliasing filters do not need to change
With “conservative” filter examples shown, WWiSE PER simulations show several dB of ACR/AACR margin for 20 MHz HT MIMO modes relative to .11a/g 54 Mbps levels
The added dispersion due to required narrower filter transition band does not significantly impact channel performance
Sean Coffey, Texas Instruments, et al.

39. New technical information - II
Short training sequence experimental results:
r0, “Short training sequence compatibility with legacy g devices,” Tushar Moorti, Christopher Young, Chris Hansen.
Analytical results show that 400 nsec shift provides best power estimation accuracy
Laboratory results show that 400 nsec shift provides the best legacy protection
Backward compatible MIMO preambles based on a 400 nsec cyclic shift on the 2nd antenna (2 TX case) are appropriate for n
Sean Coffey, Texas Instruments, et al.

40. New technical information - III
AGC impact of preamble design:
r0, “AGC power variations with WWiSE cyclic preamble structures,” Dave Hedberg, Mark Webster, Michael Seals.
The relatively long cyclic shifts employed in the WWiSE Greenfield preambles result in well behaved AGC power statistics
The AGC power statistics for mixed mode receptions of 11n signals are the same as for Greenfield modes – since the STS portion of the two preamble types is the same
The legacy portion of mixed mode preambles employ shorter cyclic shifts for the legacy LTS/SF portion – and consequently exhibit larger variation – for MM signal-field decoding
Extensive testing of legacy device detection capability with this format is documented in IEEE /1590r0
Sean Coffey, Texas Instruments, et al.

41. New technical information - IV
Pilot performance:
r0, “WWiSE pilot scheme performance,” Allert van Zelst, Richard van Nee, VK Jones.
For a 2  2 antenna configuration (which is the minimal requirement to support the mandatory modes of WWiSE (and other proposals)), the WWiSE pilot mapping outperforms the 11a pilot mapping, while having the same tracking bandwidth.
If one decides to not fully comply with the mandatory modes by, e.g., designing a 2  1 system, the 11a pilot processing can still be outperformed when the pilot processing for the WWiSE pilot mapping is averaged over two symbols.
Sean Coffey, Texas Instruments, et al.

42. New technical information - V
WWiSE preambles and beamforming:
r1, “Preambles, beamforming, and the WWiSE proposal,” Chris Hansen.
Contrary to speculative arguments in 05/1635, you can apply SVD beamforming to WWiSE preambles
Receiver processing works the same way
Excellent channel estimation performance
Sean Coffey, Texas Instruments, et al.

43. New technical information - VI
MAC mechanisms performance:
r0, “TGn features vs. performance,” Matthew Fischer.
HTP Burst performance
Block Ack/No Ack comparison
MSDU Aggregation performance
Multipoll performance loss
Sean Coffey, Texas Instruments, et al.

44. Conclusion
Sean Coffey, Texas Instruments, et al.

45. Convergence on n
The Task Group has made considerable progress towards a unified approach to n
We believe the WWiSE proposal has significant advantages, compared to other possibilities
High performance, robust and efficient core design
Full compatibility with further extensions
Modular, consistent, extendible
Minimizes testing and interoperability concerns
Minimizes time to market for a high performance, worldwide n standard
Sean Coffey, Texas Instruments, et al.

46. References and further information
IEEE / n, “WWiSE group PHY and MAC specification.”
IEEE / n, “WWiSE proposal response to functional requirements and comparison criteria.”
IEEE / r3-000n and –r4-000n, “WWiSE complete proposal presentation.”
See also
Or send to
Sean Coffey, Texas Instruments, et al.