Multi-sensor scheduling for reliable target tracking in wireless sensor networks

Due to uncertainties of target and limited sensing range and possible miss detection of sensor, single sensor based collaborative target tracking, as addressed in existing literature, suffers from the low tracking accuracy and unreliability problem when a scheduled sensor fails to detect the target. In this paper, a multi-sensor scheduling scheme for collaborative target tracking in WSNs is proposed where at each time step, a number of sensors are selected based on a specified detection probability. These tasking sensors form a temporary tracking group with one of them being the leader to carry out fusion estimation of target using measurements from the tasking sensors. The leader is also responsible for tasking sensor selection for the next time step. Due to the difficulties in, calculating the joint, detection probability of multiple sensors, a Monte Carlo method is adopted for calculating the detection probability. An incremental sensor selection heuristic is proposed to select tasking sensors to meet the specified detection probability requirement Simulation results show that, compared to a single sensor scheduling approach, the proposed approach can significantly improve the tracking accuracy and the reliability.