On the Statistical Analysis of the Multipath Propagation Model Parameters for Power Line Communications

This paper proposes a fitting procedure that aims to identify the statistical properties of the parameters that describe the most widely known multipath propagation model (MPM) used in power line communication (PLC). Firstly, the MPM parameters are computed by fitting the theoretical model to a large database of single-input-single-output (SISO) experimental measurements, carried out in typical home premises. Secondly, the determined parameters are substituted back into the MPM formulation with the aim to prove their faithfulness, thus validating the proposed computation procedure. Then, the MPM parameters properties have been evaluated. In particular, the statistical behavior is established identifying the best fitting distribution by comparing the most common distributions through the use of the likelihood function. Moreover, the relationship among the different paths is highlighted in terms of statistical correlation. The identified statistical behavior for the MPM parameters confirms the assumptions of the previous works that, however, were mostly established in an heuristic way.