A low-complexity stopping criterion for the decoder of parallel concatenated LDPC codes

In this paper, we propose a novel stopping criteria that based on the mean of the magnitude of the extrinsic information in the iterative decoder of parallel concatenated LDPC (low- density parity-check) codes. The proposed criterion has low computation complexity and few storage requirements. Simulations results proved that the MEM (mean of extrinsic information magnitude) rules outperforms the other existing criterions in terms of BER and average number of super iterations.