Markov-Middleton Impulsive Noise Mitigation in the MIMO Power Line Communication Channel

Abstract

The aim of this work is to improve the BER performance in the power line communication (PLC) channel. Impulsive noise is known to cause significant bit error in this channel. As an objective to decrease the bit error rate, adaptive filtering techniques based on the LMS and the RLS algorithms are employed for the mitigation of the impulsive noise in a MIMO PLC channel. To our best knowledge, adaptive filtering techniques employed for impulsive mitigation in MIIO PLC channels are few. Impulsive noise is represented by the Markov-Middleton and Middleton class A noise models. The Marloov-Middleton noise for a weakly and heavily disturbed MIMO PLC channel are simulated and their performances with noise mitigation are compared with that of the Middleton class A noise model. The performance of adaptive filtering techniques based on the LMS and RLS algorithms for mitigation of Markov-Middleton noise from the MIMO PLC system was remarkable. An average gain of 12.62 dB and 16.26 dB for LMS filtering, and 12.90 dB and 16.83 dB for RIS filtering were recorded in a weakly and heavily disturbed MIMO PLC channel respectively. It was concluded that for weakly disturbed MIDO PLC channels the performance of both algorithms is equally efficient while in heavily disturbed MIMO PLC channels the performance of RLS filtering slightly outperforms that of LMS filtering