Bernoulli-Gaussian Distribution with Memory as a Model for Power Line Communication Noise

The adoption of Additive Bernoulli-Gaussian Noise (ABGN) as an additive noise model for Power Line Communication (PLC) systems is the subject of this paper. ABGN is a model that for a fraction of time is a low power noise and for the remaining time is a high power (impulsive) noise. In this context, we investigate the usefulness of the ABGN as a model for the additive noise in PLC systems, using samples of noise registered during a measurement campaign. A general procedure to find the model parameters, namely, the noise power and the impulsiveness factor is presented. A strategy to assess the noise memory, using LDPC codes, is also examined and we came to the conclusion that ABGN with memory is a consistent model that can be used as for the evaluation of the noise effect on the digital communication over PLC channel even when the PLC system uses the memorydependent (i.e. error control codes) consistency that do not hold when using the (memoryless) ABGN model.