A novel self-adaptive iterative decoding scheme based on stopping criterion

This paper proposes a novel iterative decoding scheme based on stopping criterion for BICM-ID embedded turbo codes. To achieve a lower complexity, the fixed maximum and minimum iterative numbers (Imax/Imin) scheme, whose Imax and Imin are determined by the coding gains (CG) criterion, has been used. However, iterative numbers which correspond to different SNRs are varied. In order to receive better flexibility, this novel scheme is proposed as a self-adaptive decoding scheme. Both Imin and Imax, whose values are not unique, are determined by the statistics of the average iterative numbers and the CG criterion. Further improvement of the adaptability also results from different decoding schemes adopted separately in different regions of SNRs. Simulation results confirm that, compared with the traditional schemes, similar BER performance with better flexibility and much lower complexity can be achieved.