An Application Research of Kalman Filter Based Algorithms in ECEF Coordinate System for Motion Models of Sensors
暂无分享,去创建一个
Abstract Kalman filter based algorithms such as unscented Kalman filter (UKF), and the unbiased conversion measurement Kalman filter (UCMKF) are the most popular nonlinear filters which used in tracking, navigation, estimation and information fusion. Applying those filters directly in East-North-Up (ENU) coordinates with motion models of sensors causes the degradation or even divergence of filter performance. To address this issue, we first analyzed and discussed the motion model consistency of moving sensor with a constant velocity (CV). Next, we proposed to extend the application of common filter algorithms to Earth Centered Earth Fixed (ECEF) coordinates to filter random errors. We verified the validity of our proposed method by filtering random errors in a constant velocity motion model of radar. The theoretical analysis and simulation results show that the extended algorithms provide better efficiency and compatibility in moving sensors.
[1] Masao Masugi,et al. Automatic Parameter Setting Method for an Accurate Kalman Filter Tracker Using an Analytical Steady-State Performance Index , 2015, IEEE Access.
[2] Chongzhao Han,et al. Comments on "Unbiased converted measurements for tracking" , 2004 .
[3] Y. Bar-Shalom,et al. Unbiased converted measurements for tracking , 1998 .
[4] Yaakov Bar-Shalom,et al. Decorrelated unbiased converted measurement Kalman filter , 2014, IEEE Transactions on Aerospace and Electronic Systems.