在直角坐标系中采用适于大机动目标的“当前”统计模型对目标机动建模。针对目标状态模型是线性的,测量模型是非线性的情况,使用两步滤波器。对目标加速度方差和测量噪声方差进行在线估计,实现了自适应滤波。所设计的估计器与EKF估计器进行仿真比较,结果前者误差能较快收敛并且精度较高。因此所设计的估计器适于红外成像导弹拦截大机动目标。

For highly maneuvering targets, the target model was built in Cartesian coordinates based on "current" statistical model. Since target state model was linear and measurement model was nonlinear, the two-step filter was used. Adaptive filtering was a-chieved by an online estimation of target acceleration variance and measurement noise variance. For highly maneuvering targets, simulation results show that the estimator's errors can converge more rapidly and are less than EKF estimator's errors. So the tar-get state estimator is applicable to infrared imaging missile for highly maneuvering targets.