The Study And Analysis Of MAP Algorithm In Turbo Decoding

Turbo codes, first proposed by C.Berrou et.al. in 1993, are parallel concatenation of recursive systematic convolutional codes whose encoder consists of a parallel concatenation of two or more convolutional codes. It is demonstrated that Turbo codes just have 0.5db errorcomparing to the Shannon limit with BER=10-5[1] and they are not sensitive to the variationof channel. So in the last few years Turbo codes have become very important point in code field. Especially Turbo codes are extensively applied in the development of the third generation mobile communication systems[2] . Nowadays the research of Turbo codes focuses on the design of decoding algorithm, the interleaver and decoding Construction. The decoding algorithms can be dlivicled into two major classes, maximum a posteriori(MAP) and Viterbi Algorithm(VA)[3] The MAP family includes the symbol-by-symbol MAP algorithm, which is the optimal method for producing the APP information; its additive form, the Log-MAP algorithm; and its suboptimal additive form, the Max-Log-MAP algorithm. The other class of decoding algorithms is based on the modification of the VA. The SOVA family includes the soft-output Viterbi algorithm(SOVA)and the serial list output Viterbi algorithm(SLVA)[4].Until now, Turbo codes have become a hot point in communication coding for its good performance. However, the effective theory for Turbo codes should be developed. Some problems, such as sequence delaying and high computational complexity, should be resolved.