Low-Complexity Iterative Detection Based on Bi-Directional Trellis Search

Paper in proceeding, 2003

A new algorithm to reduce the computational complexity of iterative decoding is proposed. The algorithm carries out a bi-directional trellis search in a relatively smaller trellis (compared with the full-complexity trellis structure) to find two sets of sequences: the forward sequence set and the backward sequence set. The soft information is generated by considering a window of W trellis sections and the forward and the backward sequences entering it. To assess the performance, numerical results for serially concatenated convolutional codes, transmitted over an intersymbol interference channel, disturbed by additive white gaussian noise are presented. The simulation results show that for the considered system, the algorithm can obtain bit error rate performances close to that of full-complexity iterative decoding with much lower complexity.