较差质量数据下的返回弹道计算及对落点影响的分析

针对某老式卫星返回时雷达跟踪数据曾发生的测量异常及质量较差等情况，为解决此类数据在返回弹道及落点计算中出现的实际困难以及给出应对措施，基于工程实际情况给出了三种不同建模类型下的弹道计算方法——当前统计解耦扩展卡尔曼滤波（EKF）、动力学建模无迹卡尔曼滤波（UKF）、多项式滑窗最小二乘估计算法，并应用该次质量较差的实际观测数据进行了算法应用效果分析与落点计算情况对比，对计算中的经验及教训进行了总结。研究结果表明，所给出的三种算法在面对较差质量数据时性能差异较大，在工程实际应用中可考虑将多种算法互为参考。研究结果及其经验教训对卫星返回弹道及落点的实时估计的工程实施有一定的参考价值。

Reentry trajectory calculation with poor quality data and analysis of influence on impact point

Some exception and poor quality measuring data is once existed in the reentry flying of an old satellite,which made a lot of diffculities to the calculation of trajectory determination. In order to solve the problems and give some experience of using this type of data,three different calculation methods are given based on the actual situation of project,including the current statistical modeling Extended Kalman Filter(EKF),the dynamic modeling Unscented Kalman Filter(UKF) and sliding window polynomial Least Mean Square(LMS) estimation. Also the real poor quality measuring data is used to test the three methods and the trajectory calculation results and corresponding impact points are given. The experiences and lessons in the calculation are summarized. The calculation results show that the three algorithms have some obvious differences when the poor quality data is used. So a variety of algorithms are advised to provide reference for each other in a actual project application. The results and experiences have some certain reference value for reentry trajectory determination and real-time estimation of the impact point.