An improved mixture unscented Kalman filters algorithm for joint target tracking and classification
Kun Zhan, Long Xu, Hong Xu Jiang, Liang Bai, Mengjie Wu · 2014
For the joint target tracking and classification (JTC) problem with the kinematic radar only, an improved mixture unscented Kalman filters (MUKF) algorithm is proposed. The kinematic measurements and the prior speed information envelop are used to estimate the dynamic state and classify the target. Based on the traditional mixture Kalman filters (MKF) algorithm, the MUKF algorithm adopt the unscented transform (UT) to approximate the non-linear and non-Gaussian state distribution. With the improved mutual feedback strategy, our algorithm utilizes the feedback information completely and increase the tracking efficiency on the higher probable class. Mathematical analysis and simulation results confirm the better performance of the proposed method.