Fully distributed source localization via sparse reconstruction in wireless sensor network

For energy-based multi-source localization in wireless sensor networks (WSNs), multi-resolution search and expectation maximization algorithms are of either high energy consumption or low localization accuracy. In this paper, we propose a multi-source localization algorithm via sparse reconstruction based on fully distributed structure (MLSR-FD) in WSNs. The fully distributed structure is used to reduce the energy consumption of WSNs. It requires the local data transmission and processing. Each sensor node transmits energy observations to its nearest neighbors (nodes within one-hop communication range) and also receives energy observations from its nearest neighbors. In order to reduce the computational complexity, multi-resolution redundant dictionary is used to locate sources using sparse reconstruction. Finally, the robust consensus algorithm fuses the local source position estimates of the nearest nodes and converges to the global estimates. Simulation results show that the proposed MLSR-FD algorithm provides a good tradeoff between localization accuracy and energy consumption.