一种新的机动目标跟踪滤波的模型方法

本文提出了一种新的机动目标跟踪滤波的模型方法。它将目标的机动加速度作为一状态变量引入模型而直接进行估计,并通过卡尔曼滤波器的残差来检测目标机动与否。一旦检测出目标机动,马上重新启动卡尔曼滤波器以适应机动加速度的跳变。新的自适应滤波方法在这种情况下实现了最佳滤波。计算机仿真结果表明,在计算量远少于Moose方法计算量的情况下,本文方法的滤波精度与Moose方法的滤波精度相当。

A NEW MODELING FILTERING METHOD FOR TRACKING MANOEUVRING TARGETS

A new modeling filtering method for tracking manoeuvring targets is presented. The manoeuvring accelaration of target is treated as a state variable in the model and then estimated directly. The occurences of target maneuvres are detected by the residue of the Kalman filter. As soon as the manoeuvres found, the algorithm of Kalman filter is restarted so as to adapt the filter to the jumps of manoeuvring accelaration. The enw adaptive filtering method may be thought of optimum in the case. The computing time is much less than that of Moose's method while the filtering accuracy in both cases is nearly the same as ishown in the computer simulation.