An Improved Target Tracking Algorithm Based on Extended Kalman Filter for UAV

Abstract

With the sudden maneuver produced by a target, Extended Kalman Filter (EKF) will produce a non-ignored error. Then the sensor's observation range may be exceeded. In order to solve this problem, an optimization for target tracking based on EKF is proposed in this paper, exploiting sliding window to monitor the innovations to detect whether EKF has diverged and a modified EKF with expanding dimension model considering acceleration jerk is considered to be practicable to continue the algorithm instead. Simulation results are presented comparing EKF to improved EKF in different conditions and illustrating the feasibility and availability of the improved EKF.