Measurements and statistical modelling for time behaviour of power line communication impulsive noise

This paper proposes a suitable Markov chain stochastic model for the stationary and impulsive noise present in the indoor low-voltage power line communication (PLC) channel in the broadband frequency range of 1-30 MHz, based on extensive measurements. This statistical noise model has three regimes (states) representing background noise, single impulse noise and burst impulse noise. The Markov chain mainly translates the characteristics of PLC noise processes as a catenation of several successive noise impulses, which are consecutively originated in time, producing regular time sample span. When used to produce PLC noise events, the produced data sample sequence proves to conserve the original distribution of the noise measurements with considerable accuracy. In this way, it is considered for examining the time behaviour of PLC noise as a Markov chain with stochastic matrices. From this set of measured data, analytical distribution functions are proposed to model the distribution of measured values. The model enables a more accurate prediction and simulation of PLC system performance and availability.