Improved noisy gradient descent bit-flipping algorithm over Rayleigh fading channel

<span>Gradient descent bit flipping (GDBF) and its many variants have offered remarkable improvements over legacy, or modified, decoding techniques in case of low density parity check (LDPC) codes. GDBF method variants, such as noisy gradient (NGDBF) been extensively studied their performances assessed multiple channels binary symmetric channel (BSC), erasure (BEC) additive white Gaussian noise (AWGN) channel. However, performance the said decoders more realistic conditions not equally studied. An improved algorithm is proposed this paper that optimally decodes LDPC encoded codewords Rayleigh fading under various fade rates. Comparing to NGDBF method, our decoder provides substantial both error code, number iterations required achieve performance. It subsequently reduces end-to-end latency applications with ultra-low requirements.</span>