改进UKF算法及其目标跟踪性能研究  

Research on Improved UKF Algorithm and Its Target Tracking Performance

作　　者：陈伟衡[1] 赵毅寰[1] 

机构地区：[1]中国空空导弹研究院,河南洛阳471009

出　　处：《现代电子技术》2011年第23期4-6,共3页Modern Electronics Technique

摘　　要：研究了Unscented变换的基本原理及UKF算法。为了降低跟踪系统计算的复杂性,在Unscented变换中,通过引入单位矩阵,以简单的数值计算取代复杂的矩阵分解求解矩阵平方根的过程,把UKF改进为FMSRUKF。通过对三维坐标系下作变加速运动目标的跟踪仿真,结果表明FMSRUKF有更好的精度和鲁棒性。The principle of Unscented transformation and UKF（Unscented Kalman Filter） algorithm are studied.To simplify computational complexity of the tracing system,identity matrix was introduced into the Unscented transformation,and the complicated matrix calculation was replaced by simple numerical calculation for solving matrix square root,that is UKF algorithm was improved into FMSRUKF（Fixed Matrix Square Root Unscented Kalman Filter）.The simulation results of variably accelerated motion target tracing under three dimensional coordinate show that FMSRUKF achieves better precision and robust.

关 键 词：跟踪 Unscented变换 UKF FMSRUKF 

分 类 号：TN959[电子电信—信号与信息处理]