A sparse sigma-point consensus filter for wireless sensor networks

In this paper, we address a sparse signal reconstruction problem of nonlinear dynamic system in Wireless sensor networks (WSNs). A distributed reconstruction algorithm of sparse nonlinear signal is proposed based on sigma-point filter and consensus filter with embedded pseudo-measurement (PM) technology. For numerical accuracy, its square-root version is further developed. By embedding the PM equation, proposed distributed reconstruction algorithm is able to fuse far fewer random linear measurements from different nodes in the WSNs, such that sparse nonlinear signals can be reconstructed closely and all filters can reach a consensus on the estimation. The simulation results demonstrate the effectiveness of the proposed algorithm.