基于强跟踪扩展卡尔曼粒子滤波的巡航段自主导航方法

 姚文龙1,2a， 张均东1*， 刘媛2b

 （1.大连海事大学 轮机工程学院，辽宁 大连 116026；

2.青岛远洋船员职业学院 a.机电系； b.船舶与海洋工程系，山东 青岛 266071）

摘要： 通过对探测器巡航段状态方程和观测方程非线性问题的研究，提出一种强跟踪扩展卡尔曼粒子滤波（STEPF）算法，并将其应用于巡航段自主光学导航的方案中，由此来实时确定探测器的轨道。所提出的改进粒子滤波算法是将强跟踪扩展卡尔曼滤波引入粒子滤波来更新粒子，产生重要性密度,缓解粒子退化和样本贫化问题，以提高导航系统对状态突变的跟踪能力。通过仿真表明，改进粒子滤波算法在探测器轨道参数精度和方差预报的有效性方面有了较大地提高，能够适应快速变化的飞行环境，而且滤波精度明显的优于EPF滤波算法。

关键词： 自主导航； 强跟踪扩展卡尔曼粒子滤波； 粒子滤波； 探测器； 巡航段

中图分类号： V 448.2      文献标志码： A

Autonomous Navigation of Cruise Phase Based on Strong TrackingEPF

YAO Wenlong1,2a, ZHANG Jundong1, LIU Yuan2b

(1.School of Marine Engineering, Dalian Maritime University, Dalian 116026, China

2.a.Department of Marine Engineering； b.Department of Naval Architecture and  Ocean Engineering,Qingdao Ocean Shipping Mariners College, Qingdao  266071，China)

Abstract: Aiming at the nonlinearity of state equation and observation equation, an improved particle filter algorithm based on strong tracking extent Kalman filter (STEKF) was presented, and it was imported into the autonomous optical navigation scheme of cruise phase, the real time orbit for probes was determined later. The strong tracking filter was used to update particles in the new algorithm and produce importance densities. As a result, the problems of particle degeneracy and sample impoverishment were ameliorated, and the tracking ability was improved. Simulation result shown that the strong tracking extent particle filter algorithm (STEPF) was not only improving the prediction accuracy of the orbital parameters and deviation, but also sloving the problem of the orbit determination in the rapid flight environment. It had the better accuracy than EPF algorithm.

Key words: autonomous navigation；strong trackingEPF； particle filter；spacecraft；cruise phase

收稿日期： 20140331

基金项目： 国家自然科学基金项目（51179102）；山东省高等学校科技计划项目（J13LN72）；青岛市市南区科技发展资金项目（20125004ZH）.

作者简介： 姚文龙(1981—)，男，博士研究生.     *通信联系人.