8/10/2019 Robust Mobile http://slidepdf.com/reader/full/robust-mobile 1/15 1 Robust Mobile Multimedia Conferencing Robust Mobile Multimedia Conferencing The Third IEEE Workshop on Wireless Local Area The Third IEEE Workshop on Wireless Local Area Networks Networks Thursday, September 27 Thursday, September 27 th th , 2001 , 2001 Izzet Agoren Izzet Agoren Center for Information & Center for Information & Communications Technology Communications Technology Research Research Department of Electrical Department of Electrical Engineering Engineering The Pennsyl vania State University, The Pennsyl vania State University, University Park, PA University Park, PA Mohsen Mohsen Kavehrad Kavehrad
•• Devise and incorporate effective error resiliency techniques intDevise and incorporate effective error resiliency techniques into theo theexisting low bit rate video coding standard, H.263. These technexisting low bit rate video coding standard, H.263. These techniquesiques
will protect against both packet (group of blocks (GOB)) loss anwill protect against both packet (group of blocks (GOB)) loss and lossesd losses
due to individual bit errors. Error due to individual bit errors. Error --resiliency requires augmentations ofresiliency requires augmentations of
both the source encoder and the decoder.both the source encoder and the decoder.
– – General Work StrategyGeneral Work Strategy
•• Work from existing (C) source code for H.263 (from Telenor)Work from existing (C) source code for H.263 (from Telenor)
•• Identifying insertion points and “ least invasive” methodology foIdentifying insertion points and “ least invasive” methodology forr
implementation of error implementation of error --resiliency featuresresiliency features
•• Debugging and Performance EvaluationDebugging and Performance Evaluation
Robust Mobile Multimedia ConferencingRobust Mobile Multimedia Conferencing
•• Error Resilient H.263 Codec TechniquesError Resilient H.263 Codec Techniques – – Error Recovery at the Decoder Error Recovery at the Decoder
•• Al ternat ive motion vector select ion for lost Intermode blocks Al ternat ive motion vector selection for lost Intermode blocks•• Block refresh for Intramode blocksBlock refresh for Intramode blocks
– – RateRate--distor tion Optimized Inter/Intra MB Mode Selection At thedistor tion Optimized Inter/Intra MB Mode Selection At theEncoder, Consistent With Decoder Error RecoveryEncoder, Consistent With Decoder Error Recovery
•• Inter Inter --coding of MBs requires far fewer bits ( ~ 1/3) than those requir coding of MBs requires far fewer bits ( ~ 1/3) than those requir ededfor Intrafor Intra--coding.coding.
•• However, under random packet loss, intermode coding incurs muchHowever, under random packet loss, intermode coding incurs muchgreater expected distor tion, with errors propagated through t imegreater expected distor tion, with errors propagated through t ime..
•• These facts motivate rateThese facts mot ivate rate--distortion optimized MB mode selection.distortion optimized MB mode selection.
•• Expected distort ion measured consistent with decoder's error recExpected distort ion measured consistent with decoder's error recovery.overy.
•• Estimation takes into account error propagation (recursively estEstimation takes into account error propagation (recursively estimateimatefor current frame based on estimate for past f rame).for current frame based on estimate for past frame).
•• Expected distortion evaluated at pixelExpected distortion evaluated at pixel--level precision .level precision .
•• Mode selection for each MB to minimize the Lagrangian cost D(modMode selection for each MB to minimize the Lagrangian cost D(mode) +e) +Lambda R(mode)Lambda R(mode)
•• Devise and incorporate smart antenna strategies which ut ilize muDevise and incorporate smart antenna strategies which ut ilize mult ilt i--path diversity of uplink and downlink wireless channels. The smapath diversity of uplink and downlink wireless channels. The smartrtantenna will reduce coantenna will reduce co--channel interference at base station andchannel interference at base station andincrease signal power at mobile station, which wi ll result in loincrease signal power at mobile station, which wil l result in lower bitwer biterror rate and higher video quality.error rate and higher video quality.
– – General Work StrategyGeneral Work Strategy
•• Use Signal Processing Worksystem (SPW™), a blockUse Signal Processing Worksystem (SPW™), a block--oriented, physicaloriented, physical
layer endlayer end--toto --end simulation tool, which conforms to current standardsend simulation tool, which conforms to current standardsfor 3G WCDMA. Design and build blocks that implement smart antenfor 3G WCDMA. Design and build blocks that implement smart antennanafeatures. Incorporate H.263 codec to evaluate overall systemfeatures. Incorporate H.263 codec to evaluate overall systemperformance.performance.
•• [1][1] -- A. A. RavindranRavindran et al., "Mobile Multimedia Services for Third Generation Communet al., "Mobile Multimedia Services for Third Generation Commun icationsicationsSystems,Systems,”” IEEE VTC, 2001.IEEE VTC, 2001.
•• [2][2] -- I.I. HabbabHabbab, M., M. KavehradKavehrad & C. E.& C. E. SundbergSundberg, "ALOHA with capture over slow and fast fading, "ALOHA with capture over slow and fast fadingchannels with coding and diversity, " IEEE JSAC, vol. 7, pp. 79channels with coding and diversity, " IEEE JSAC, vol. 7, pp. 79--88, Jan 1989.88, Jan 1989.
•• [3][3] -- Video coding for low bit rate communication, ITUVideo coding for low bit rate communication, ITU--T recommendation H.263 version 2, 1998.T recommendation H.263 version 2, 1998.
•• [5][5] – – R. Zhang, L.R. Zhang, L. RegunathanRegunathan, and K Rose,, and K Rose, ““ Video coding with optimal Inter/IntraVideo coding with optimal Inter/Intra--modemodeswitching for packet loss resilience,switching for packet loss resilience,”” IEEE Journal on Selected Areas inIEEE Journal on Selected Areas in CommunicatiosCommunicatios ,, vo lvol 18,18,no. 6, pp. 966no. 6, pp. 966--976, 2000.976, 2000.
•• [6][6] – – B.B. WidrowWidrow and J. Stearns, Adaptive Signal Processing. Prenticeand J. Stearns, Adaptive Signal Processing. Prentice--Hall, 1985Hall, 1985
•• [7][7] – – Cynthia E. Anderson and Mark A.Cynthia E. Anderson and Mark A. WickertWickert ,, ““ The Performance of a Wireless LAN AccessThe Performance of a Wireless LAN AccessNode Using AntennaNode Using Antenna BeamformingBeamforming for Dynamic and Static Users,for Dynamic and Static Users,”” IEEE 1999IEEE 1999
•• [8][8] – – M. R.M. R. SturgillSturgill , G. Cortez, R., G. Cortez, R. Avinun Avinun , and S. M., and S. M. Alamouti Alamouti,, ““ Design and Verification of ThirdDesign and Verification of Third
•• [9][9] – – M. R.M. R. SturgillSturgill , and S. M., and S. M. Alamouti Alamouti ,, ““ Simulator Simulator --based design methodology for Wirelessbased design methodology for WirelessCommunications Systems,Communications Systems,”” Tech. Rep., Cadence Design Systems, Inc., December 2000Tech. Rep., Cadence Design Systems, Inc., December 2000
•• [10][10] – – J.J. FeiginFeigin, K., K. PahlavanPahlavan & M.& M. YlianttilaYlianttila,, ““ HardwareHardware--FittedFitted ModellingModelling and Simulation of and Simulation of VoIPVoIPover a Wireless LAN,over a Wireless LAN,”” IEEE VTC 2000IEEE VTC 2000
•• [11][11] – – M.M. FreytesFreytes, C. E., C. E. RodrigiguezRodrigiguez & C. A. Marques,& C. A. Marques, ““ RealReal--time H.263+ Videotime H.263+ Video TransmissioTransmissio onon802.11 Wireless LANs,802.11 Wireless LANs,”” IEEE 2001IEEE 2001
