Unscented Kalman Filter Target Tracking Based on Elman Neural Network

Abstract

Most of target motion models have uncertainties and strong nonlinear characteristics, in the uncertainty processing, the effect of Unscented Kalman Filter (UKF) is remarkable. In order to achieve accurate target tracking, based on the UKF algorithm, this paper uses the ability of dynamic and non-linear approximation of Elman neural network to compensate filtering results of UKF. In addition, this paper explains the shortcomings of Extended Kalman Filter(EKF) algorithm in strong non-linear systems from mechanism. By tracking targets in different motion situations, the emulation results indicate that the algorithm which has proposed in this paper is superior to the traditional UKF and EKF algorithm from filtering accuracy and error stability.