Abstract:

The problem of multi-sensor multi-task scheduling for target tracking tasks is studied. In view of the mobility of targets, the extended Kalman filter method for target tracking is adopted. A mixed integral programming model is founded, in which three contradictory objectives, the overall priority for scheduling tasks successfully, the accuracy of target tracking and the consumption of energy in the sensor network, are considered. The grouping particle swarm optimization algorithm is then introduced, which avoids too fast convergence and prematurity by partitioning the swarm into groups and enabling each group evolving toward its own best position. Numerical results show that the presented method is efficient in solving the sensor scheduling problem.