卫星姿态确定的分布式非线性滤波方法  

Decentralized nonlinear filters for spacecraft attitude estimation

作　　者：范春石[1] 张高飞[1] 孙剑[1] 孟子阳[1] 尤政[1] 

机构地区：[1]清华大学精密仪器与机械学系,精密测试技术及仪器国家重点实验室,北京100084

出　　处：《清华大学学报（自然科学版）》2010年第2期224-228,共5页Journal of Tsinghua University(Science and Technology)

基　　金：国家“八六三”高技术项目(2005AA714012)

摘　　要：针对轻型星敏感器、微机电系统(MEMS)陀螺、太阳敏感器和磁强计构成的轻型化卫星姿态确定系统,设计了一种分布式非线性融合滤波结构,提出一种快速采样点姿态估计算法,仅需4个采样点,即可实现完整的姿态及陀螺漂移估计,精度不低于常规采样点滤波,而运算量显著降低,与扩展Kalman滤波(EKF)算法相当。综合考虑器件特点,采用方差插值,融合子滤波器的相关信息,实现异类传感器的优势互补,从而获得高精度、高实时性、持续稳健的姿态及速率估计。仿真结果验证了算法的有效性。This paper presents a decentralized satellite attitude/rate estimation system instrumented with a suit of micro sensors including an active pixel sensor （APS） star tracker, micro-electro-mechanical system （MEMS） gyroscopes, an APS sun sensor, and MEMS magnetometers. A reduced sigma point Kalman filtering algorithm using only 4 sigma points is used to give complete attitude and rate estimates. The algorithm＇s computational efficiency is significantly improved to rival its extend Kalman filter （EKF） counterpart, while still insuring 2nd-order precision as in traditional sigma point Kalman filters. Real time quaternion and gyro-bias states of the local filters are fused by covariance intersection to obtain precise and robust attitude and rate estimate, while the weakness of a single sensor is ameliorated to some extent. Simulation results verify the algorithm＇s efficiency.

关 键 词：姿态确定 分布武滤波 采样点Kalman滤波 非线性 方差插值 

分 类 号：V412.42[航空宇航科学与技术—航空宇航推进理论与工程]