Learning the Impedance Entanglement for Wireline Data Communication

In wireline communication networks, a line impedance entanglement exists since changes of the line impedance at one network port cause a change of the line impedance at the other port. This physical phenomenon can be constructively exploited to realize a form of digital modulation that is referred to as impedance modulation (IM). IM is an alternative method to more conventional voltage modulation (VM). In this paper, the impedance entanglement is studied and learned through a supervised machine learning (ML) approach which enables the implementation of a ML based receiver. Numerical results are obtained in a data set of measured power line communication channels, which is among the most challenging environments for such a modulation approach. The resulting system can have practical implementation, for instance in a smart building automation network where monitoring-control of sensors and devices enables the efficient energy management. Comparisons with the optimal maximum-likelihood (MaxL) receiver that perfectly knows the impedance entanglement transfer function are made. It is found that the ML based receiver performs close to the optimal genie receiver.