电视跟踪系统非线性测量延迟滤波算法研究

Kalman Filtering with Nonlinear Measurement Delay in TV Tracking Systems

针对电视跟踪系统存在的脱靶量现象所导致的不能直接进行目标运动参数预测及预测精度降低问题,提出在测量方程中引入一不确定因子γ以表明系统存在测量延迟,且由γ的概率来表示系统的延迟量。同时根据不敏卡尔曼滤波(UKF)原理,给出带有测量延迟的非线性系统的状态估计计算方法(称其为测量延迟UKF算法)。并在电视跟踪伺服控制系统中对此算法进行仿真实验。实验结果证明此算法在电视跟踪等效复合控制系统中的有效性,且实验结果也表明在测量方程中引入一不确定因子γ以表明系统存在测量延迟是有效可行的,而且此算法可大幅度降低速度预测误差和跟踪误差。

Considering the problems that the target motion parameters can not be predicted directly and the predicting accuracy is decreased due to the miss distance in TV tracking system,an uncertain factor γ was introduced in the measurement equation,and the probability of γ was used to represent the system delay.According to the theory of Unscented Kalman Filtering（UKF）,the calculation algorithm of state estimation of the nonlinear system with measurement delay was given,which was then simulated in equivalent compound control of TV tracking system.The simulation results proved the effectiveness of the algorithm in the equivalent compound control,and the experimental result also showed it is feasible to introduce an uncertain factor in measurement equation for describing measurement delay,which can decrease predicting and tracking error greatly.