Sparsity-promoting sensor selection for nonlinear target tracking with quantized data
Xiaojun Yang, Ruixin Niu · 2017
In this paper, sparsity-promoting sensor selection algorithms for target tracking with quantized data are developed. We formulate sensor selection as an optimization problem that aims to strike a balance between estimation accuracy and the number of selected sensors. To cope with sensor selection problems in large-scale wireless sensor networks (WSNs), we propose a fast centralized optimization algorithm based on the alternating direction method of multipliers (ADMM). We further develop a low-complexity distributed version of the ADMM where each sensor makes a local sensor selection decision. The simulation results show that the proposed centralized and distributed algorithms activate the most informative sensors and yield very good tradeoff between the estimation performance and the cost of sensing and communication. For large scale sensor networks, the distributed ADMM algorithm is more efficient and has lower computational load per sensor node.