摘要
针对一类未知机动目标设计了一种线性最优滤波器。该滤波器采用有限记忆模型作为滤波模型,模型参数用动态最小二乘方法辨识得到。由此得到的滤波器称为基于辨识有限记忆模型的线性最优滤波器。有限记忆模型克服了时间多项式模型(如CV模型、CA模型、Singer模型等)对目标物理特性的依赖性,它从目标运动轨迹入手对目标运动特性进行描述,更适合于战场未知目标的跟踪滤波。通过原理分析和Monte Carlo仿真实验验证了该滤波方法的有效性,并且和Kalman滤波方法进行了比较。
An optimal linear filter is designed for a class of targets with unknown maneuvers. The model of the filter is Finite Memory Model （FMM） with identified parameters. Dynamic Least Square method is used to identify these parameters. The filter designed is named as the Optimal Linear Filter based on Finite Memory Model with Identified Parameters. Not as the time-polynomial model （such as CV model, CA model, Singer Model etc. ）, the FMM is independent on the physical character of the target. It describes the kinetic character of the target from its track, which makes the FMM is more suitable for the filtering and tracking of these unknown targets on the battlefield. The principle of modeling of FMM is analyzed in this paper, and the filter＇s validity is verified by the Monte Carlo simulation experiments. In these experiments the filter proposed is compared with several kinds of Kalman filters.
出处
《火力与指挥控制》
CSCD
北大核心
2007年第9期95-100,共6页
Fire Control & Command Control
基金
国防预研基金资助项目
