A.N. Lutkov - PJSC «NPO Almaz» (Moscow). E-mail: anlutkov@gmail.com

Application of the Block Turbo Code or Turbo Product Code in discrete messages - transmission system is one of the best methods for solution the problem of the noise-proof coding. It is most actual for systems which transmit single-messages of small size (100−200 symbols) at the low signal-to-noise ratio. The correcting ability of the Block Turbo Code is closely to the limits dictated by channel ca-pacity and the decoding complexity is suitable for implementation in many applications. Finding decoding algorithms with reduced complexity for real time systems is actual today. This article deals with Chase-Pyndiah turbo-decoding algorithm. The one may be considered as the reference at the product code de-coding with minimizing block error probability. The method to simplify the algorithm is investigated. Simplification is reached by means of reducing the revaluated symbols number in rows and columns of the code matrix. Performance characteristics of the initial decoder are plotted for different number of the revaluated symbols. Accumulation of the extrinsic information is proposed for increase the tolerance from reducing the number of revaluated symbols. The better law of the accumulation is established. Analytical justification for the modified block diagram of the decoder with extrinsic information store is performed. According to the modified block diagram additional operations for the accumulation are not required. Characteristics of the coding gain losses relatively to the initial algorithm are plotted for fixed PB = 10-4 at variation of the revaluated symbols number and restriction of the maximum iterations number. The ability of the decoder-s complexity reducing is shown. Simplification of the proposed algorithm with extrinsic information store is reached by means of reducing as the maximum iteration count and as the revaluated symbols number. The proposed algorithm with extrinsic information store allows for reduce decoding complexity on 17−28% with losses near 0,1−0,2 dB respectively.

 

1. Zubarev JU.B., Ovechkin G.V. Pomekhoustojjchivoe kodirovanie v cifrovykh sistemakh peredachi dannykh // EHlektrosvjaz. 2008. № 12. S. 58−61.
2. Channelcodingincommunicationnetworks: fromtheorytoturbocodes/ EditedbyAlainGlavieux. 1sted. ISTELtd. 2007.
3. Vargauzin V. Vblizi granicy SHennona // Telemultimedia. Ijun. 2005. S. 3−10.
4. Arkhipkin A. Turbokody - moshhnye algoritmy dlja sovremennykh sistem svjazi // Besprovodnye tekhnologii. 2006. № 1. S. 63−64.
5. Goalic A., Pyndiah R.Real-TimeTurbo-DecodingofProductCodesonaDigitalSignalProcessor // IEEE. 1997. P. 624−628. 0-7803-4198-8/97.
6. Elias P.Error-FreeCoding // IERTrans. onInf. Theory. September. 1954. V. IT‑4. P. 29−37.
7. Pyndiah R., Glavieux A., Picart A.,Jacq S.NearOptimumDecodingofProductCodes // IEEE. 1994. P. 339−343.
8. Pyndiah R.IterativeDecodingofProductCodes // ProceedingsoftheInternationalSymposiumonTurboCodes. Brest. France. September 1997. P. 71−79.
9. Pyndiah R.Nearoptimumdecodingofproductcodes: BlockTurboCodes // IEEETrans. onCommunications. August 1998. V. 46. P. 1003−1010.
10. Gladkikh A.A. Osnovy teorii mjagkogo dekodirovanija izbytochnykh kodov v stirajushhem kanale svjazi. Uljanovsk: UlGTU. 2010. 379 s.