基于UKF的实时弹道解算参数稳健设计

摘要

In realtime traj ectory estimation based on Unscented Kalman Filter (UKF),the influence of unstable ex-ternal environment and existence of outliers made filtering results inaccurate or even divergent. Therefore,it was necessary to adj ust and determine values of filter parameters properly to overcome the problem before a missile launch. A robust design of filter parameters based on theoretical traj ectory as prior information was proposed to pre-dict the optimal value of parameters. Starting from the perspective of filter framework,this paper selected relevant parameters,adopted the idea of robust design,constructed a ballistic measurement model through simulating trajec-tory feature points based on theoretical traj ectory in outer design;and carried out orthogonal experimental design by reversing measurement data from constructed model,then j udged the significance of the parameters and determined its optimal value,so as to give guidance to parameter setting of realtime traj ectory estimation. Examples show that this method is of high validity,and can improve the robustness and convergence performance of the filter.%在基于UKF(Unscented Kalman Filter,无迹卡尔曼滤波)的实时弹道解算应用中,外部环境不稳定的影响以及观测过程中野值的存在,使得基于经验参数设置的滤波易于发散.因此,需要针对指定型号调节和设置滤波器参数的取值,使其克服上述现象.针对这一问题,提出了基于理论弹道先验的参数稳健设计方法,对滤波器参数最优值进行预测.从滤波器的架构出发,梳理出影响滤波器性能的相关参数,结合稳健设计思想,外设计部分基于理论弹道先验信息,通过模拟弹道特征点构造弹道观测模型;而内设计部份,构造模型的测元数据进行正交试验,判断参数的显著性,获取相应的最优参数取值,从而为实时解算弹道的参数设置提供指导.实际数据试验结果表明,该方法可显著提高滤波器的实算稳健性和收敛性能.