采用自适应无迹卡尔曼滤波的卫星姿态确定

针对现有算法卫星姿态确定中模型参数估计不准确,系统存在外界干扰下稳定性差和跟踪精度不足的问题,提出一种自适应无迹卡尔曼滤波算法,对卫星三轴姿态进行估计。首先分析了陀螺和星敏组合定姿的工作原理,然后推导了以误差四元数为状态变量的卫星姿态运动学方程。滤波过程中,该算法引入自适应矩阵,对量测噪声协方差矩阵进行调整;依据滤波发散判别准则,对系统噪声协方差矩阵进行自适应修正,抑制滤波过程中可能的发散情形,获得了良好的自适应性能。实验结果表明,在参数估计不准确时,自适应无迹卡尔曼滤波相比鲁棒自适应UKF算法,三轴估计精度的均方根误差(RMSE)分别提升了30.0%,34.1%,22.4%。该算法基本满足卫星姿态确定的高精度、强鲁棒性等要求。

Attitude determination for satellite using adaptive unscented Kalman filter

Poor stability and low tracking accuracy are significant issues in existing algorithms for satellite attitude determination.An adaptive unscented Kalman filter(AUKF)algorithm was proposed to over?come these issues and estimate the three-axis attitude of satellite by modeling error and external distur?bance.First,the working principle of attitude determination based on gyro sensor was analyzed,following which the satellite attitude kinematics equation,with error quaternion as state variable,was derived.An adaptive matrix was introduced to adjust the measurement noise covariance matrix.Based on the filtering divergence criterion,the system noise covariance matrix was adaptively modified to suppress potential di?vergence in the filtering process,and a good adaptive performance was obtained.Finally,it is demonstrated through experimental verification that,compared with robust AUKF algorithm,the accuracy of threeaxis estimation(RMSE)of AUKF improves by 30.0%,34.1%,and 22.4%,respectively,when the pa?rameter estimation is not accurate.Thus,the algorithm meets the requirements of high precision and strong robustness for satellite attitude determination.