不完全量测下光电跟踪系统估计误差分析

Estimation Error Analysis on Optic-electric Tracking System with Intermittent Observations

为了评估跟踪系统在不完全量测下的跟踪性能,该文利用投影定理给出了修正的去偏转换量测卡尔曼滤波(UCMKF)算法,进一步利用修正的黎卡提方程与修正的克拉美-罗下界(CRLB)给出了跟踪系统统计意义下估计误差协方差一致有界的充分条件,以及跟踪系统统计意义下CRLB的一组递推上下界。最后,Monte-Carlo仿真表明:随着采样时间的增加,递推上下界逐渐逼近了跟踪系统统计意义下CRLB的枚举真值。

In order to evaluate the tracking performance of tracking systems with intermittent observations, a modified unbiased converted measurement Kalman Filtering （UCMKF） algorithm is presented by means of projection theorem. A sufficient condition for uniform boundedness of the average of estimation error covarianee, and a group of recursive upper and lower bounds of the average Cramer-Rao low bounds （CRLB） of tracking systems are obtained using a modified Riecati equation and a modified CRLB. Monte-Carlo simulation results show that the recursive upper and lower bounds are close to the average CRLB of tracking systems as the sample-collecting time increases.