基于修正的卡尔曼滤波自适应跟踪算法  

An Adaptive Tracking Algorithm Based on Kalman Filter

作　　者：赵易峰[1] 李京华[1] 彭京晶[1] 许家栋[1] 

机构地区：[1]西北工业大学电子信息学院,陕西西安720072

出　　处：《计算机仿真》2008年第8期59-63,共5页Computer Simulation

基　　金：国防科技重点实验室基金项目(51454070204HK0320)

摘　　要：针对卡尔曼滤波对匀速运动目标能有效的跟踪,但是当目标出现转弯时,很难达到跟踪精度的要求,甚至丢失目标的现象。对卡尔曼滤波算法进行了改进,在观测向量中引入了两个加速度误差变量,它们动态地修正状态估计误差从而减少跟踪精度误差,形成了修正的Kalman算法。但是由于状态变量维数增加,使得计算量增加,实时性下降,将卡尔曼滤波算法与修正的卡尔曼滤波算法两种算法相结合,提出了基于修正的卡尔曼滤波自适应跟踪算法。仿真结果表明,具有良好的稳定性和精确度,优于一般的卡尔曼滤波算法。Concerning the problem of the low accuracy, even lost the target when the Kalman filter is applied under the target swerving condition, the paper mended the Kalman filter and introduced two error variables of acceleration for modifiying the state estimation,then generated a new modified Kalman algorithm. But because the state variable dimension increases, the computation quantity also correspondingly increases. In order to satisfy the timeliness, the paper improved the adaptive tracking algorithm by combining both the Kalman filter algorithm and modified Kalman filter algorithm. The simulation result shows that the adaptive tracking algorithm based on Kaiman filter has a favorable stability and accuracy.

关 键 词：卡尔曼滤波 自适应跟踪 加权衰减因子 协方差 

分 类 号：TN913.36[电子电信—通信与信息系统]