Markov Model Characterization of a Multicarrier Narrowband Powerline Channel With Memory in an Underground Mining Environment

The error distribution of multicarrier modulation in a narrowband powerline communication (PLC) channel with memory is presented for an underground mining environment. In the environment, the noise in the PLC channel originates from the mains switchboard and a blast control unit, connected to the powerline link. Using the error vectors measured from the channel, the memory channel model is obtained by training the measured data using a hidden Markov model and the Fritchman model for channel state classification. The channel with memory is modeled by considering the state transition probabilities between the current state and one previous state of the channel. The measured and modeled data are then compared in order to determine the suitability of the derived models. Using the error-free run distribution and probabilities of error of the modeled data, the modeled data match the measured data, validating the suitability of the derived models for an underground mining environment.