Enhanced Performance of Narrowband Power Line Communications using Recursive Least Squares Filter

Abstract

Noises presented in a power line communication channel tend to distort the message signals, leading towards the reception of erroneous data at the receiver end. Mitigation of noise existent in power line has always been of prime interest and helped to improve the BER performance of a communication system as it accounts for efficient data transmission. In this work, adaptive filters based on the Recursive Least Squares (RLS) algorithm and Least Mean Square (LMS) algorithms have been implemented in Simulink to investigate the effectiveness of an Adaptive Noise Canceller for the mitigation of Gaussian and Impulsive Noises present in a narrowband power line channel model. The performance of the RLS algorithm against that of the LMS algorithm was compared in adaptive filtering for the same channel conditions. The error performances of BPSK and FSK schemes for the channel model in a generic digital communication system were also compared in Simulink. Furthermore, the use of convolutional codes and interleaving for the correction of random bit errors and for condensing the negative effect of burst errors respectively were investigated during the transmission of data signals over the generic communication system designed in Simulink. From the findings of the study, it has been concluded that the RLS algorithm proves to be more effective than the LMS algorithm. For a BER of 10-5, a coding gain of less than 10 dB is achievable for both Binary Phase Shift Keying and Binary Frequency Shift Keying. With the addition of convolutional coding and convolutional interleaving, the error performance of the channel is further improved, rendering the power line channel more reliable for data communication.