SER Analysis of OFDM-Based Power-Line Communications under the Impulsive Nature of Alpha-Stable Background Noise

In an orthogonal frequency-division multiplexing (OFDM)-based power-line communication (PLC) system, the background noise is commonly assumed as Gaussian. However this model may not truly depict the effect of the human activities on noise characteristics, leading to system performance overestimation. Plenty of empirical evidences show that the impulsive nature of the PLC background noise can be well modeled by a symmetric alpha-stable (SαS) process. In this paper, we investigate the analytical symbol error rate (SER) performance of M-ary modulation schemes for OFDM-based PLC systems subject to SαS background noise. The strength of the SαS noise is evaluated by geometric signal-to-noise ratio (GSNR) based on the zero-order statistics. We conduct extensive simulation experiments and numerical analyses, showing that under SαS background noise, the fast Fourier transform (FFT) of the received complex baseband SαS noise samples follow another bivariate SαS distribution, and the SER performance of the OFDM system decreases as the employed FFT size grows.