The design of a soft decision variable rate sequential decoder using a modified stack algorithm is described. The feedback systematic convolutional code is used for encoding, and a syndrome decoding method is used for decoding. The stack decoder searches the most likely paths using the syndrome sequence, which is formed by a syndrome former. A dummy-bit-insertion technique permits effective decoding of high-rate codes, such as R=3/4 and 7/8. It is concluded that since this stack decoder provides a coding gain much higher than the conventional Viterbi decoder, its application to power-limited satellite or space communication systems will make it possible to use transmission power more efficiently and/or to improve the transmission quality. >

Development of a variable rate syndrome sequential decoder based on a stack algorithm

An analysis of the BER performance of sequential decoding with stack algorithm is presented for rate 1/2 code. A tight upper bound of BER performance for maximum likelihood decoding is calculated by specific code distance profiles for evaluating the proto error probability of the simulation results. Block error probability is also considered through the random coding theory so that the simulation results satisfy its upper bound. Finally, the coding gain achievable by the sequential decoder is presented for given constraint length of the code and the memory capacity in the decoder, which will be useful for the hardware design.

Yasuo Hirata

Papers

Estimation of bit error rate performance of sequential decoding