In view of the fact the accuracy of the third-degree Cubature Kalman Filter (CKF) used for initial alignment under large misalignment angle conditions is insufficient, an improved fifth-degree CKF algorithm is proposed in this paper. In order to make full use of the innovation on filtering, the innovation covariance matrix is calculated recursively by an innovative sequence with an exponent fading factor. Then a new adaptive error covariance matrix scaling algorithm is proposed. The Singular Value Decomposition (SVD) method is used for improving the numerical stability of the fifth-degree CKF in this paper. In order to avoid the overshoot caused by excessive scaling of error covariance matrix during the convergence stage, the scaling scheme is terminated when the gradient of azimuth reaches the maximum. The experimental results show that the improved algorithm has better alignment accuracy with large misalignment angles than the traditional algorithm.
For the strong nonlinearity of initial alignment with large misalignment angles of Strapdown Inertial Navigation System (SINS), improving the accuracy of the nonlinear filtering algorithm is important for navigation. In order to solve the non-local sampling problem of Cubature Kalman Filter (CKF) algorithm when the state dimension increases, the orthogonal transformation of sigma points is introduced to reduce the high-order component errors of the first moment of CKF. A short-term adaptive algorithm with better robustness is proposed on the basis of adaptive filtering algorithm and Transformed CKF (TCKF), also making better convergence efficiency and accuracy of initial alignment with large misalignment angles. For the application of non-linear initial alignment, the influence of the sub-optimal algorithm of adaptive filter on the stability of alignment is reduced by controlling the number of effective filtering cycles of adaptive alignment. To summarise, there are two main motivations for this paper. Firstly, utilizing innovation in adaptive methods to accelerate the convergence speed of initial alignment. Secondly, by using statistical methods to find deterministic short-term correction strategies, the accuracy and robustness of alignment can be improved under the premise of convergence. The effectiveness of the improved algorithms is verified by simulation and turntable experiments.
The paper first analyzes the impact of nonlocal sampling problem on the high order estimation errors of sigma point filter. Then, we introduce that how TCKF reduces the impact of nonlocal sampling on filtering by introducing orthogonal matrices to sigma points generating framework, thus improving the estimation accuracy of high order terms. In order to further reduce the high order estimation errors of CKF, an orthogonal matrix cluster and PSO optimization algorithm are introduced in this paper. By minimizing the coefficients of the 4-th order terms of the Taylor series of state estimation, the optimized orthogonal matrix is obtained, and then the improved TCKF algorithm is proposed. The effectiveness of the algorithm is demonstrated through simulation experiments and turntable initial alignment experiment. The method proposed in this paper has good generality for 2nd order and above sigma point filtering algorithm.