基于四元数的UKF多传感器卫星姿态确定  

Satellite Attitude Estimation with Multisensor Based on Quaternion Using UKF

作　　者：陈志明[1] 王惠南[1] 刘海颖[1] 

机构地区：[1]南京航空航天大学自动化学院,江苏南京210016

出　　处：《上海航天》2012年第2期13-17,共5页Aerospace Shanghai

摘　　要：对采用磁强计和太阳敏感器的卫星姿态模型,应用无迹卡尔曼滤波(UKF)算法确定卫星姿态。由四元数描述姿态,在算法中将其转换成旋转矢量,另设计了四元数求均值的迭代算法,解决了四元数在UKF算法中的处理。仿真结果表明:广义卡尔曼滤波(EKF)算法对初值依赖性较大,UKF算法不受初值影响且精度高于EKF,具快速性、精度高和稳定性强等优点。Attitude based on quaternion was estimated by unscented Kalman filtering（UKF） arithmetic for a satellite model with magnetometers and sun sensors in this paper.The quaternion was transferred into rotation vector in order to match the arithmetic,and the algorithm was designed to calculate the mean quaternion to solve the treatment of the quaternion in UKF.The simulation results showed that UKF algorithm was unaffected by initial data and more effective with high precision and stabilization,while EKF algorithm was relied on the initial condition very much.

关 键 词：卫星姿态测量 四元数 无迹卡尔曼滤波 矢量观测 

分 类 号：V448.22[航空宇航科学与技术—飞行器设计]