An improved non-compressed beamforming feedback format for 11ac

Slides:



Advertisements
Similar presentations
Doc.: IEEE /1645r2 Submission January 2005 C. Hansen, BroadcomSlide 1 Preambles, Beamforming, and the WWiSE Proposal Notice: This document has.
Advertisements

Doc.: IEEE /0935r1 Submission July 2011 Fei Tong, CSRSlide 1 An improved non-compressed beamforming feedback format for 11ac Date:
Beacon Measurement on Pilot Frames
Reference waveform generator for 11ac: Status Update
Submission on comments to +HTC frames
LB84 General AdHoc Group Sept. Closing TGn Motions
LB84 General AdHoc Group Sept. Closing TGn Motions
[ Interim Meetings 2006] Date: Authors: July 2005
Motions Date: Authors: January 2006
London TGu Motions Authors: January 2007 Date: Month Year
LB73 Noise and Location Categories
LB73 Noise and Location Categories
Waveform Generator Source Code
Preambles, Beamforming, and the WWiSE Proposal
March 2014 Election Results
TGp Closing Report Date: Authors: July 2007 Month Year
[ Policies and Procedure Summary]
[ Policies and Procedure Summary]
Motion to accept Draft p 2.0
3GPP liaison report July 2006
[place presentation subject title text here]
Motions Date: Authors: January 2006
TGp Motions Date: Authors: November 2005 Month Year
TGp Closing Report Date: Authors: March 2006 Month Year
On Coexistence Mechanisms
TGu-changes-from-d0-02-to-d0-03
TGp Closing Report Date: Authors: May 2007 Month Year
TGn Frame Format Ad Hoc Status and Motions
On Coexistence Mechanisms
TGp Closing Report Date: Authors: March 2006 Month Year
Reflector Tutorial Date: Authors: July 2006 Month Year
TGv Redline D0.07 Insert and Deletion
TGv Redline D0.06 Insert and Deletion
July 2012 Opening Report Date: Authors: July 2012
Scaling and Quantization for CSI Matrices Feedback
TGu-changes-from-d0-01-to-d0-02
LB73 Noise and Location Categories
TGy draft 2.0 with changebars from draft 1.0
Suggested Resolution on CID #10074
TGv Redline D0.10 Insert and Deletion
WAPI Position Paper Sept 2005 Sept 2005 IEEE WG
Redline of draft P802.11w D2.2 Date: Authors:
TGr Proposed Draft Revision Notice
Off-channel selection
TGu-changes-from-d0-02-to-d0-03
[ Policies and Procedure Summary]
Preambles, Beamforming, and the WWiSE Proposal
TGu Motions Date: Authors: May 2006 May 2006
Liaison Report From Date: Authors: Month Year
Beamforming and Link Adaptation Motions
Reference waveform generator for 11ac
[ Policies and Procedure Summary]
Benefits of Compressed Explicit feedback to resolve CID 2969
Draft P802.11s D1.03 WordConversion
Questions to the Contention-based Protocol (CBP) Study Group
January Opening Report
Suggested Resolution on CID #10074
Reference waveform generator for 11ac
EC Motions – July 2005 Plenary
TGu-changes-from-d0-04-to-d0-05
STC with CSI feedback IEEE P Wireless LANs Date:
TGu-changes-from-d0-03-to-d0-04
TGu Motions Date: Authors: May 2006 May 2006
PSMP Adhoc Oct TGn Adhoc
Beamforming and Link Adaptation Motions for LB 84 Comment Resolutions
Explicit feedback with sounding packet
WAPI Position Paper Sept 2005 Sept 2005 IEEE WG
TGr Proposed Draft Revision Notice
TGp Motions Date: Authors: January 2006 Month Year
May 2012 Opening Report Date: Authors: May 2012
Presentation transcript:

An improved non-compressed beamforming feedback format for 11ac July 2011 doc.: IEEE 802.15465651-11/0xxxr0 July 14, 2011 An improved non-compressed beamforming feedback format for 11ac Date: 2011-07-14 Authors: Notice: This document has been prepared to assist IEEE 802.11. 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 802.11. Patent Policy and Procedures: The contributor is familiar with the IEEE 802 Patent Policy and Procedures <http:// 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 <stuart.kerry@philips.com> 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 802.11 Working Group. If you have questions, contact the IEEE Patent Committee Administrator at <patcom@ieee.org>. Fei Tong, CSR Fei Tong, CSR

An improved non-compressed beamforming feedback format for 11ac July 14, 2011 An improved non-compressed beamforming feedback format for 11ac CSR Fei Tong, CSR

Summary Propose an improved non-compressed beamforming feedback format July 14, 2011 Summary Propose an improved non-compressed beamforming feedback format Demonstrate that, for equal performance, non-compressed BF offers similar or lower feedback overhead (bits per matrix) to compressed BF, for 1 or 2-antenna beamformee devices Slide 3 Fei Tong, CSR

Focus on beamforming feedback with one or two column vectors July 14, 2011 Focus on beamforming feedback with one or two column vectors Beamforming matrix has one or two column vectors The only option in devices with one or two Rx antennas Feedback matrix dimension is small : (Ntx, 1) or (Ntx, 2) Complexity of BF matrix calculation is low Why? A large proportion of portable devices with embedded 802.11 modules will have one or two Rx antennas due to physical size and power consumption constraints For middle to long propagation distances, 1 or 2 spatial streams is still preferable even if there are >= 2 Rx antennas for reliable communication and option to reject interference Fei Tong, CSR

Improved non-compressed feedback format July 14, 2011 Improved non-compressed feedback format Each column vector is normalised by its element with maximum magnitude before quantisation Beamforming vector is norm and phase invariant Benefits Can maximise the quantisation resolution by choosing different scaling for each beamforming vector Reduced bit count with same quantisation accuracy: only feedback (Nt-1) complex values and one index to represent the position of the maximal magnitude in each vector 1 bits for this index for 2 Tx antennas 2 bits for this index for 3 and 4 Tx antennas 3 bits for this index for 5, 6, 7 and 8 Tx antennas Fei Tong, CSR

Quantisation accuracy and overhead study July 14, 2011 Quantisation accuracy and overhead study Accuracy Compare the CDF of quantisation error (in dBc) over a set of beamforming matrices (8x1 or 8x2) obtained from IEEE802.11 fading channel B random realisations Overhead The total number of bits needed with a given accuracy to feedback a beamforming matrix Fei Tong, CSR

Nt x 1 BF matrix, lowest resolution July 14, 2011 Nt x 1 BF matrix, lowest resolution Compressed format (Feedback type = 0, Codebook Information = 0) (2,4) Non-compressed format with Nb=2 uses fewer bits than the compressed format (2,4) Slide 7 Fei Tong, CSR

Nt x 2 BF matrix, lowest resolution July 14, 2011 Nt x 2 BF matrix, lowest resolution For greater than 3 Tx antennas, non-compressed format with Nb=2 requires fewer bits than compressed format (2,4). Slide 8 Fei Tong, CSR

Nt x 1 BF vector, medium resolution July 14, 2011 Nt x 1 BF vector, medium resolution Compressed format (Feedback type = 0, Codebook Information = 1) (4,6) Non-compressed format with Nb=4 uses fewer bits than compressed format (4,6) Slide 9 Fei Tong, CSR

Nt x 2 BF vector, medium resolution July 14, 2011 Nt x 2 BF vector, medium resolution For greater than 3 Tx antennas, non-compressed format with Nb=4 requires fewer bits than compressed format (4,6). Slide 10 Fei Tong, CSR

Nt x 1 BF vector, highest resolution July 14, 2011 Nt x 1 BF vector, highest resolution Compressed format (Feedback type = 1, Codebook Information = 1) (7,9) Non-compressed format with Nb=6 uses fewer bits than compressed format (7,9) Slide 11 Fei Tong, CSR

Nt x 2 BF vector, highest resolution July 14, 2011 Nt x 2 BF vector, highest resolution For greater than 3 Tx antenna, non-compressed format with Nb=6 uses fewer bits than compressed format (7,9) Slide 12 Fei Tong, CSR

SU-MIMO performance simulation setting July 14, 2011 SU-MIMO performance simulation setting SU-MIMO mode using IEEE Fading channel model B No channel variation due to feedback delay Equal 64 QAM, BCC 2/3 coding rate, 125 Bytes PPDU MMSE receiver with channel estimation SVD on estimate channel and quantised feedback (FB) Beamforming based on FB; number of streams equals to number of column vector in FB; equal power on each stream Non-compressed FB: 2 or 3 Bits quantisation per real value Compressed FB: low resolution feedback format (2,4) Fei Tong, CSR

Frame error rate results for 1 SS (1) July 14, 2011 Frame error rate results for 1 SS (1) Total = 5 bits per tone for Non-comp Nb=2 Total = 7 bits per tone for Non-comp Nb=3 Total = 6 bits per tone for Compressed format Total = 14 bits per tone for Non-comp Nb=2 Total = 20 bits per tone for Non-comp Nb=3 Total = 18 bits per tone for Compressed format Fei Tong, CSR

Frame error rate results for 1 SS (2) July 14, 2011 Frame error rate results for 1 SS (2) Total = 23 bits per tone for Non-comp Nb=2 Total = 33 bits per tone for Non-comp Nb=3 Total = 30 bits per tone for Compressed format Total = 31 bits per tone for Non-comp Nb=2 Total = 45 bits per tone for Non-comp Nb=3 Total = 42 bits per tone for Compressed format Fei Tong, CSR

Frame error rate results for 2 SS (1) July 14, 2011 Frame error rate results for 2 SS (1) Total = 10 bits per tone for Non-comp Nb=2 Total = 14 bits per tone for Non-comp Nb=3 Total = 6 bits per tone for Compressed format Total = 28 bits per tone for Non-comp Nb=2 Total = 40 bits per tone for Non-comp Nb=3 Total = 30 bits per tone for Compressed format Fei Tong, CSR

Frame error rate results for 2 SS (2) July 14, 2011 Frame error rate results for 2 SS (2) Total = 46 bits per tone for Non-comp Nb=2 Total = 66 bits per tone for Non-comp Nb=3 Total = 54 bits per tone for Compressed format Total = 62 bits per tone for Non-comp Nb=2 Total = 90 bits per tone for Non-comp Nb=3 Total = 78 bits per tone for Compressed format Fei Tong, CSR

Complexity of compression July 14, 2011 Complexity of compression Counts of number of complex multiplication and arctan/arccos for one BF matrix The number of arctan/arccos is quite large Fei Tong, CSR

MU-MIMO performance simulation setting July 14, 2011 MU-MIMO performance simulation setting MU-MIMO mode using IEEE Fading channel model B No channel variation due to feedback delay 2 Rx antenna per user, total 2 users Each user feedback the vector corresponding to the strongest spatial stream ZF beamforming based on FB vector from 2 users Equal 64 QAM, BCC 2/3 coding rate, 125 Bytes PPDU for all users MMSE receiver with channel estimation per user Fei Tong, CSR

FER results for 1 SS (Nt x 1 FB) July 14, 2011 FER results for 1 SS (Nt x 1 FB) Total = 9 bits per tone per user for Non-comp Nb=4 Total = 12 bits per tone per user for Compressed format (5,7) Total = 26 bits per tone per user for Non-comp Nb=4 Total = 36 bits per tone for Compressed format Fei Tong, CSR

July 14, 2011 Conclusions For 1 or 2 column BF matrix, non-compressed feedback has similar or lower overhead (with all quantisation resolution level) than compressed feedback with similar BF performance No extra complexity to beamformee introduced by compressing process Including non-compressed feedback format does not require changes to BF frame exchange procedure Fei Tong, CSR

Only minor changes needed to the draft July 14, 2011 Only minor changes needed to the draft Extra bit is needed to indicate the feedback format in VHT MIMO control field; reserved bits are available Extra bit is needed to indicate the compressed feedback format is supported in VHT capabilities field; reserved bits are available Fei Tong, CSR

July 2011 doc.: IEEE 802.15465651-11/0xxxr0 July 14, 2011 Strawpoll 1 Do you support including non-compressed format for low rank feedback (number of streams no greater than 2) Yes No Abstain Fei Tong, CSR Fei Tong, CSR

July 2011 doc.: IEEE 802.15465651-11/0xxxr0 July 14, 2011 Strawpoll 2 Specify that the non-compressed feedback format support is mandatory at the beamformer and compressed feedback is optional at the beamformee for Nss=1 and 2. For Nss>2, compressed feedback format remains the only option Yes No Abstain Fei Tong, CSR Fei Tong, CSR

July 2011 doc.: IEEE 802.15465651-11/0xxxr0 July 14, 2011 Strawpoll 3 Modify the 11n non-compressed format: before quantisation, the beamforming matrices are right-hand multiplied by a diagonal matrix so that the largest element of each column vector is one. For each column vector, only feedback (Nt-1) complex values and index of the element equal to one. Yes No Abstain Fei Tong, CSR Fei Tong, CSR