Topology reconstruction for power line network based on Bayesian compressed sensing

Abstract

Power line communication (PLC) is playing a more and more important role in the smart grid (SG). In this paper, in addition to conveying information, sensing topology of the network based on PLC is proposed to expand the area of PLC applications, which further improves the smart property of the grid. By assuming each endpoint of the grid is equipped with a PLC device, we model the grid as an edge-node network with tree structure. Then, considering the parametric sparsity of the PLC channel, we propose a method to estimate the distances between nodes using Bayesian compressed sensing (CS). Finally, we exploit the proposed dynamic reconstruction algorithm to reacquire the topology of the whole network. Numerical simulation results demonstrate that the proposed Bayesian CS scheme can accurately achieve the distances between nodes especially with fewer number of pilots, while the dynamic reconstruction algorithm is more effective and has less complexity than the traditional method.