强跟踪自适应滤波器实现机动目标的精确跟踪

机动目标状态估计中的一个主要问题是:目标运动的突变性导致状态噪声无法进行统计预测.传统的EKF将噪声看成是高斯白噪声有着本质上的不足,因而无法实现稳定的跟踪.引入Sage-Husa滤波算法对有色噪声进行在线的估计,一定程度上弥补了目标运动模型不够合理的缺憾.在此基础上,从系统容错设计基本原理出发,用归一化残差功率法实时地检测可能出现的数值发散现象,一旦检测到发散,印通过一种改进的强跟踪自适应滤波器进行抑制,有效地提升了滤波的健硕性,实现了稳定跟踪.最后,针对高机动目标的运动特性,仿真验证采用变维滤波模型,用EKF对目标的简单机动进行跟踪,只有目标运动突变时才采用本文提出的算法,以提升计算的实时性.仿真结果表明此算法对高机动目标的跟踪是有效的.

Maneuvering target tracking based on a strong tracking adaptive filter

It is a main problem in maneuvering target tracking that the process noise statistics is unknown due to the maneuverability of the target.It is an inherent shortcoming of traditional EKF to regard this noise as white noise,which can not ensur stable tracking.So,the Sage-Husa noise estimation algorithm is adopted to estimate the colored noise on line,which can compensate the defect of the faulty model to some degree.Based on which,based on the principle of fault tolerance design in systems,the innovation var...