Adaptive thresholding algorithm for detection of coded signals in narrowband powerline communication systems

Abstract

Powerline Communication (PLC) signals are affected by a combination of different sources of noise. Among them, impulse noise represents the most harmful component, making the use of encoding algorithms an essential requirement. Algorithms based on Complementary Sequences (CS) are an attractive option for signal detection in conditions of low signal-to-noise ratios (SNR). One of the main characteristics of CS is the high SNR gain by computing the sum of their autocorrelation functions, which, in a noiseless scenario, results in a Kronecker delta. In the presence of noise, the amplitude of main deltas is no longer proportional to the length of the sequences, and a variable number of random amplitude sidelobes emerge around it. This work analyses some of the algorithms present in the bibliography and proposes a novel dynamic algorithm to detect the correlation deltas immersed in the PLC impulse noise. The proposal is based on the estimation of the noise variance and the correlation peaks in order to compute a near-optimal validation threshold. The proposal is simulated and evaluated in a PLC environment with a Middleton Class A noise model, and the algorithm performance is compared in a real communication test over the electrical network of the university building.