Decoding algorithm of LDPC codes for cascaded BSC-AWGN channels

The written-in errors in bit patterned media (BPM) recording systems cause the erroneous bits during the writing process, leading to performance degradation. In previous works, the BPM read/write channel is modeled as a binary symmetric channel (BSC) with an additive white Gaussian noise (AWGN) channel or the cascaded BSC-AWGN channels. The initial channel information including the written-in errors probability has been proposed for low-density parity check (LDPC) decoder. However, the decoding algorithm remains the same as in the case of AWGN channels. In this work, we show that the crossover probability of BSC channels will affect the check node processing of LDPC codes. Therefore, we modify the check node computation that takes into account of the crossover probability of BSC channels. The simulation results show that the proposed LDPC decoder outperforms the ones with the conventional and modified likelihood function in terms of the bit error rate performance.