摘要
传统算法在解决纯方位目标跟踪时存在有偏、收敛速度慢或发散等不足,无迹卡尔曼滤波(UKF)虽然改善了系统线性化误差,但并没有明显改善卡尔曼滤波器容易发散的问题。文章在扩展卡尔曼滤波和UKF算法的基础上,提出一种衰减记忆UKF算法(MAUKF),引进衰减因子加强对当前测量数据的利用,减小历史数据对滤波的影响。理论分析和仿真结果表明,MAUKF算法在纯方位目标跟踪中的滤波精度、稳定性和收敛时间都优于EKF、UKF算法。
The traditional algorithms applied in bearings-only target tracking have some shortages or disadvantages such as biased, slow convergence or divergence. The UKF algorithm improves the linearization of system, but it doesn＇t amend the robustness of system obviously. In this paper, an improved UKF named MAUKF （Memory Attenuated Unscented Kalman Filtering） algorithm which is based on the extended Kalman filter and the UKF is proposed. The MAUKF algorithm improves the robustness by using a fading factor. Theoretical analysis and simulation result indicate that the UKF has better performance than EKF and UKF algorithms in precision, stability and convergence time when it is applied in bearing-only target tracking.
出处
《舰船电子工程》
2009年第8期75-78,共4页
Ship Electronic Engineering
关键词
纯方位
非线性滤波
扩展卡尔曼滤波
无迹卡尔曼滤波
bearings-only, nonlinear filtering, exlended Kalman filter, unscented Kalman filter
