论文部分内容阅读
在航天器姿态确定时,针对由于模型误差而使滤波发散的情况,根据协方差匹配的方法,提出一种基于K-矩阵残差的自适应姿态确定算法。通过调整滤波器中的噪声协方差,得到一种在线最佳姿态四元数的最小二乘拟合递归算法。针对传统残差定义的不足,根据K-矩阵的几何性质,设计适应的残差和标准,来解决这一问题。针对高值的残差可由测量噪声或过程噪声的建模误差引起,或由两个共同引起,分别对测量噪声自适应滤波和过程噪声自适应滤波进行了研究。仿真表明,此种算法效果较好。
According to the method of covariance matching, an adaptive attitude determination algorithm based on K-matrix residual is proposed to solve the problem that the filter disperses due to model error when spacecraft attitude is determined. By adjusting the noise covariance in the filter, a least-squares fitting recursive algorithm of quaternion of online best posture is obtained. Aiming at the shortcomings of the traditional definition of residuals, this problem is solved according to the geometric properties of K-matrix and the design of adaptive residuals and criteria. Residuals for high values can be caused by modeling noise of measurement noise or process noise, or caused by two together. Adaptive measurement noise filtering and process noise adaptive filtering are respectively studied. Simulation shows that this algorithm works better.