用一种自适应UKF滤波器跟踪机动目标

Maneuvering Target Tracking with an Adaptive UKF Filter

在非线性系统中用自适应Kalman滤波器跟踪机动目标时,必须采用扩展Kalman形式,但扩展Kalman有精度低、易发散的缺点。在标准UKF滤波器的基础上,应用无迹变换将测量坐标系下的滤波残差统计特性转换到三维笛卡儿坐标系下,根据三维坐标下的滤波残差变化实时改变过程噪声协方差矩阵的大小,实现了自适应UKF滤波。滤波器既具备UKF滤波器精度高,不易发散的优点,同时不需了解目标机动的先验信息。仿真表明了本滤波器的有效性。

When the adaptive Kalman filter is applied to track Maneuvering targets in nonlinear system,it may replace the Kalman filter with Extend Kalman Filter(EKF),but the EKF provides bad accuracy and sometimes diverges.Based on the normal UKF filter,An adaptive UKF filter is proposed,which uses Unscented Transformation(UT) to transform the covariance of innovation from measurement Coordinates to Cartesian Coordinates,and modifies the scale factors of process noise covariance according to the change of innovation ...