A Novel Node Selection Method of Bearings-only Sensors for Target Tracking in Wireless Sensor Networks

In this paper, we propose a novel information utility description for bearings-only sensors to meet the challenge of sensor arrangement in Wireless Sensor Networks (WSN) for tracking. In order to reduce calculation, we use the probability distribution function of the predicted target state, the sensing model and the position information of sensors to define information utility, thus avoiding the computationally burdensome estimation of the posteriori distribution. Through simulations, we compare the performances of different sensor selection methods in the means of mean square error (MSE). The results show that our method can attain a good tracking accuracy while saving large amounts of calculation.