Generalized targets detection and counting in dense wireless sensors networks

This paper applies the Compressed Sensing (CS) the targets detection in small scale dense Wireless Sensors Networks (WSN). The monitored area is partitioned into cells, each equipped by one sensor. The CS application aims to locate targets from a reduced subset of sensors measurements. A generalized version of a recently proposed Greedy Matching Pursuit algorithm (GMP), designed for point events joint detection and counting, is derived, which is denoted by gGMP. This generalization enables the identification of several active cells at each iteration. Also, an optimized deterministic sensors subset selection scheme, based on the maximum energy is envisaged and shown to outperform the random choice scheme.