Syndrome-based scheduling: simple adaptive scheduling technique for decoding LDPC codes

摘要

A very simple adaptive scheduling method is introduced to improve the convergence rate of low-density parity-check decoders. Although the standard decoding method does not impose any constraint on the process order of check nodes, the proposed scheme adopts a dynamic scheduling method for check node message updating based on syndrome values. The proposed technique is applied to lower-density parity check (LDPC) codes of the 802.11n and 802.15.3c standards. Compared with sequential scheduling, the proposed approach provides a faster convergence speed from 28 to 51% for a WER of 10 and 10, respectively. Consequently, the introduced adaptive scheduling achieves a target performance with less number of iterations, which leads to reduced power consumption with minimal additional decoding complexity.