e-space
Manchester Metropolitan University's Research Repository

On Companding and Optimization of OFDM Signals for Mitigating Impulsive Noise in Power-line Communication Systems

Anoh, K and Adebisi, B and Rabie, KM and Hammoudeh, M and Gacanin, H (2017) On Companding and Optimization of OFDM Signals for Mitigating Impulsive Noise in Power-line Communication Systems. IEEE Access, 5. pp. 21818-21830. ISSN 2169-3536

[img]
Preview

Download (689kB) | Preview

Abstract

Generally, the probability density function (PDF) of orthogonal frequency division multiplexing (OFDM) signal amplitudes follow the Rayleigh distribution, thus, it is difficult to correctly predict the existence of impulsive noise (IN) in powerline communication (PLC) systems. Compressing and expanding the amplitudes of some of these OFDM signals, usually referred to as companding, is a peak-to-average power ratio (PAPR) reduction technique that distorts the amplitudes of OFDM signals towards a uniform distribution. We suggest its application in PLC systems such as IEEE 1901 powerline standard (which uses OFDM) to reduce the impacts of IN. This is because the PLC channel picks up impulsive interference that the conventional OFDM driver cannot combat. We explore, therefore, five widely used companding schemes that convert the OFDM signal amplitude distribution to uniform distribution to avail the mitigation of IN in PLC system receivers by blanking, clipping and their hybrid (clipping-blanking). We also apply nonlinear optimization search to find the optimal mitigation thresholds and results show significant improvement in the output signal-to-noise ratio (SNR) for all companding transforms considered of up to 4 dB SNR gain. It follows that the conventional PDF leads to false IN detection which diminishes the output SNR when any of the above three nonlinear memoryless mitigation schemes is applied.

Impact and Reach

Statistics

Downloads
Activity Overview
8Downloads
30Hits

Additional statistics for this dataset are available via IRStats2.

Altmetric

Actions (login required)

Edit Item Edit Item