基于UPF的航天器自主天文导航方法  

Spacecraft Autonomous Celestial Navigation Based on the Unscented Particle Filter

作　　者：宋利芳[1] 房建成[1] 

机构地区：[1]北京航空航天大学第五研究室,100083

出　　处：《航天控制》2005年第6期31-34,40,共5页Aerospace Control

基　　金：国家自然科学基金(60174031);民用航天预研项目资助

摘　　要：利用红外地球敏感器和星敏感器直接敏感地平的天文导航方法是一种成熟、可靠的自主导航方法。这种导航方法的状态方程和量测方程都是严重非线性的,且在建立航天器轨道动力学模型时,通常将二阶带谐摄动项建模,而将其他摄动项等效为高斯白噪声,由于这些摄动项都有其精确的模型,通常不服从高斯分布。本文提出将UPF(Unscented Particle Filter)滤波方法应用于航天器自主天文导航,该方法用UKF(Unscented Kalman Filter)得到粒子滤波的重要性采样密度函数,从而克服了标准的粒子滤波没有考虑最新量测信息和UKF只能应用于噪声为高斯分布的不足。仿真结果表明,该方法可以取得比标准的粒子滤波和UKF更快的滤波收敛性和更高的滤波精度。Starlight angular distance observed by star sensor and horizontal instrument as the observation is often used for the spacecraft celestial navigation, which is a mature and credible autonomous navigation method. When modeling the orbit motion, only the 2nd zonal harmonic of the non-spherical earth gravity is usually introduced to the system state equation and the other perturbation terms are regarded as white noise. But all these terms have certain model themselves, which are usually non-Gaussian distribution. Unscented particle filter（ UPF） is proposed to be used for the spacecraft celestial navigation system. The UPF uses the unscented Kalman filter（ UKF） to generate sophisticated proposal distributions. It can not only avoid the limitation of the UKF which only apply to Gaussian distributions but also avoid the limitation of the standard PF which can not include the new measurements. The simulation results show that the presented method has faster convergence rate and higher accuracy than the standard PF and UKF.

关 键 词：航天器 自主导航 天文导航 UPF 

分 类 号：V448[航空宇航科学与技术—飞行器设计]