The Analysis Of Turbo Decoder Based On Stochastic Computation

With the development of modern communications technology, the data transferspeed has been greatly improved. High throughput become the requirement for modernchannel decoder designs. Turbo code is one of the good codes, which approaches theShannon limit. Thus，Turbo code has been widely used in the areas of wireless andmobile communications with its unique advantages，and also is the most importantchannel code scheme for the third generation mobile communication system and itslong term evolution and fourth-generation mobile communication system. The variousdecoding algorithm and architecture are proposed based on the sequential decodingalgorithm and the iterations between the two SISO decoder，This method can greatlylimit the speed of the turbo decoder. In this thesis, we propose a fully parallel Turbodecoding algorithm and structure based on stochastic computation，which is based onthe full parallel extrinsic information exchanging. We provide the analysis for thefully parallel Turbo decoding algorithm and propose the architecture of the full paralleldecoder. The performance of the parallel Turbo decoding algorithm is also studied inthis thesis. The proposed decoder has several advantages, including high speed, simplearchitecture and memory free. From the viewpoint of information theory, the proposedscheme can high throughput and efficient without performance losing, compared withthe traditional decoding algorithm and realization methods. Thus,the proposed schemecan be used for the future high-speed, broadband wireless/mobile communicationssystems.