LDPC-COFDM for PLC in Non-Gaussian Noise Using LLRs Derived from Effective Noise PDFs

In this study, the performance of irregular low-density parity check (LDPC) coded orthogonal frequency division multiplexing (COFDM) utilising 4096 quadrature amplitude modulation (4096-QAM) is investigated over multipath power-line communication (PLC) channel. The effective complex-valued ratio distributions of the noise samples at the zero-forcing equaliser output considering both frequency-selective multipath PLCs, background and impulsive noise are derived, in addition to the condition for optimum detection of 4096-QAM and the bit error rate (BER). Moreover, the performance of the LDPC decoder is improved by computing the log-likelihood ratios (LLRs) required for soft decoding based on the derived probability density functions. Numerical results obtained using the newly derived LLRs demonstrate a significant performance improvement compared to the conventional receiver that uses blanking impulsive noise mitigation method and LLR computed based on the Gaussian distribution. Furthermore, EXtrinsic Information Transfer chart analysis demonstrates that the proposed approach requires fewer iterations for convergence compared to the conventional receiver. Finally, utilising channel bandwidth of 22.4 MHz, the proposed system offers an improvement of 111 Mbps over the conventional system to reach a maximum data throughput of 256 Mbps for a signal-to-noise ratio of 39 dB and BER of 10 -5 .