基于改进UKF的空间非合作目标相对导航方法  

Modified Unscented Kalman Filter for Relative Navigation of Non-cooperative Target

作　　者：张艾 程瑞 郭兰杰 侯帅 于飞[1] 李婧[1] 鄢南兴 ZHANG Ai;CHENG Rui;GUO Lanjie;HOU Shuai;YU Fei;LI Jing;YAN Nanxing(Beijing Institute of Space Mechanics&Electricity,Beijing 100094,China)

机构地区：[1]北京空间机电研究所,北京100094

出　　处：《航天返回与遥感》2022年第1期129-138,共10页Spacecraft Recovery & Remote Sensing

基　　金：国家重大科技专项工程。

摘　　要：目前,我国大力提升高精度探测指向机动能力,并发展成像跟踪能力的趋势已经形成。对机动目标的持续观测能力,良好的动中成像性能是光学遥感技术发展的重要趋势。文章针对高轨空间卫星的高精度观测与实时跟瞄需求,研究空间非合作目标的高精度相对导航问题。为了解决无迹滤波(Unscented Kalman Filter,UKF)在系统能观度较低时估计精度下降的问题,提出了一种基于能观度的改进无迹卡尔曼滤波算法(Modified Unscented Kalman Filter,MUKF),该算法定义了一种基于滤波过程误差增益阵的系统能观度表征方法,提出了基于该方法的能观度缩放参数,在线调整滤波增益协方差矩阵,使算法可根据当前时刻的能观度大小在线调整状态预测和系统观测的权重。数学仿真表明,与UKF相比,MUKF估计精度提升约4倍,且MUKF趋稳更快,精度更高。In recent years,the trend of improving the mobility of high-precision pointing and imaging tracking has been formed.The continuous observation and good imaging performance of maneuvering targets are important trends in the development of optical remote sensing technology.Considering the requirement of GEO satellites high accuracy observation and real time tracking,the relative navigation algorithm of non-cooperative target is studied in this paper.A modified unscented Kalman filter(MUKF)is proposed with a new characteristic variable formed with the system observable degree.In MUKF,the error covariance matrix from the observability theory of a nonlinear system and the weight of state prediction and system observation are adjusted on line.The numerical simulations show that the MUKF is less sensitive to the observable degree than the usual one with high accuracy.The steady state error of MUKF is about 24%of that of UKF.

关 键 词：目标监视 相对导航 无迹卡尔曼滤波 空间态势感知 

分 类 号：V448.224[航空宇航科学与技术—飞行器设计]