By using evolution of information entropy in target detection and prediction and updating matrices of covariance in Kalman filter equations
we can solve information gain in a measurement for target detection or tracking
and present a method of sensor management based on maximization of information gain.Simulation results show that this method can rationally assign the limited sensor resources and improve the whole performance of tracking system compared with sequence tracking method
and can maximally detect and track new targets under a certain tracking precision compared with random detection and tracking.