Performance analysis for decoding LT codes over BIAWGN channels with SNR mismatch
Lei Yuan, Jie Pan, Yuan Lin · 2017
In a binary input additive white Gaussian noise (BIAWGN) channel, belief propagation (BP) decoding for luby transform (LT) codes requires the knowledge of the signal-to-noise ratio (SNR) at the receiver to achieve its optimal performance. An erroneous estimation of the SNR at the decoder is referred to as “SNR mismatch”. SNR mismatch can significantly degrade the BP decoding performance of LT codes. In this paper, we propose a semi-Gaussian approximation method to predict and compare the asymptotic performance of LT codes under SNR mismatch. Asymptotic analysis and simulation results show that LT codes are more sensitive to SNR over-estimation and under-estimation for small and large SNR offset, respectively. We also show that the improved LT codes which shape the degree distribution of information nodes to lower error floor are more sensitive to SNR mismatch than original LT codes.