Modeling of amplitude-correlated and occurrence-dependent impulsive noise for power line communication

The existing Bernoulli-Gaussian (BG) model describes the impulsive noise samples for power line communication (PLQ as independent and identically distributed (i.i.d.). In practice, impulses often occur in bursts and have memory. In this paper, the impulsive noise for PLC is investigated using both measurement and modeling approaches. To the best of our knowledge, this is the first work to model the correlation between consecutive impulses for PLC, using a hybrid model which includes the weighted contributions of deterministic and random patterns of individual impulses in a burst. The occurrence dependence between bursts and between impulses in a burst is described by an enhanced two-level Markov Chain (MQ based model. An intensive performance analysis of the proposed models is provided. The effectiveness of the models is verified by measured results.