On the SCL Decoding of Polar Codes in Power Line Communications

Polar codes have been proven very powerful over additive white Gaussian noise (AWGN) channels even with short block lengths. In this paper, we consider the use of polar codes to reduce the effect of impulsive noise encountered in power line communication (PLC) channels. In particular, we consider the successive cancellation list (SCL) decoding of polar codes in the presence of Middleton class-A noise. The channel log-likelihood ratio (LLR) is derived under the two-term approximation for the Middleton class-A noise, which can be efficiently computed. Two pre-processing methods, including blanking and clipping, are investigated for possible mitigation of impulsive noise before the LLR computation. Extensive experiments show that the proposed LLR computation, along with blanking preprocessing, works well for SCL decoding of polar codes over PLC channels.