Ellipsoidal State Estimation for UAV Real Time Target Tracking

Due to the equivalent relation between UAV target tracking in practice and state estimation in academy, this paper concentrates on state estimation and its application in UAV target tracking. To relax the strict assumption on white noise for classical Kalman filter, we consider the more general bounded noise, being included in an ellipsoid. Then for better understanding our proposed ellipsoidal state estimation, three continuous processes are shown sequently, i.e. ellipsoidal approximation of arithmetic sum, ellipsoidal approximations of intersections between ellipsoid and strip, real time recursive form for generating a sequence of ellipsoids, including each state estimation at each time instant. To combine the theoretic result and engineering application, the detailed simulation example is given to prove the efficiency of our real time ellipsoidal state estimation algorithm.