论文部分内容阅读
针对由星敏感器和光学导航相机组成的卫星天文自主导航系统,传统的平方根UKF不能很好地解决测量噪声为有色噪声情况下的非线性滤波问题,导致导航系统的精度下降.为此,提出了一种有色噪声情况下的平方根UKF方法.同时,为了避免在数值计算的过程中,由于舍入误差而破坏误差协方差矩阵的正定性和对称性,在整个递推计算过程中,借鉴平方根Kalman滤波理论,采用协方差矩阵平方根进行递推计算,改善滤波算法的稳定性,协方差矩阵的平方根更新用cholesky分解和qr分解来计算.将该方法应用于卫星天文自主导航系统中,实验仿真结果表明,相对于传统的平方根UKF算法,所设计的平方根UKF算法能够很好地解决测量噪声为有色噪声情况下估计精度低问题.
For the satellite astronomic autonomous navigation system composed of star sensor and optical navigation camera, the traditional square root UKF can not solve the problem of non-linear filtering when the measurement noise is colored noise, which leads to the decrease of the accuracy of the navigation system. In order to avoid the positive definiteness and symmetry of the error covariance matrix under the condition of colored noise, in order to avoid the rounding error, the square root Kalman filter theory, the square root of covariance matrix is used for recursive computation to improve the stability of the filtering algorithm, and the square root update of covariance matrix is calculated by cholesky decomposition and qr decomposition.The method is applied to satellite astronomical autonomous navigation system, experimental simulation The results show that, compared with the traditional square root UKF algorithm, the designed square root UKF algorithm can well solve the problem of low estimation accuracy when the measurement noise is colored noise.