Low-complexity belief-propagation decoding via dynamic silent-variable-node-free scheduling

This letter presents a low-complexity scheduling scheme, referred to as the dynamic silent-variable-node-free scheduling (D-SVNFS) scheme, for the sequential belief propagation decoding of LDPC code. The D-SVNFS regulates the dynamic propagation of message updates based on the check-to-variable message residuals. To determine the next message update, it computes the on-demand message residuals and selects the largest one associated with the last updated variable and check nodes. In addition, the D-SVNFS attempts to propagate more message updates toward the erroneous variable nodes, trying to correct them with higher priority. It is shown that the proposed scheduling reduces the BP decoding complexity by up to 70% compared with prior-art SVNF scheme, without affecting the error-rate performance, at medium to high signal-to-noise ratio over the BI-AWGN and Rayleigh fading channels.