Novel strong tracking square‐root cubature kalman filter for GNSS/INS integrated navigation system

Abstract

In the GNSS/INS integrated navigation system, the accuracy of square‐root cubature Kalman filter (SCKF) will be reduced when the process model is not precisely known. To solve this problem, a novel strong tracking SCKF is proposed. The proposed algorithm utilises the measurement update module based on singular value decomposition (SVD), which can reduce the computational complexity. Moreover, a novel method to obtain suboptimal fading factor based on the orthogonality of the innovation is deduced here, which is simple to calculate and no need to solve the Jacobian matrix. This novel method introduces the calculated suboptimal fading factor into the square‐root of the state prediction covariance matrix. Then, the gain matrix can be adjusted online which can improve the robustness of the algorithm when the process model is uncertain. Meanwhile, the proposed algorithm adopts the hypothesis testing to detect the uncertainty of the process model. Then, the proposed approach will only use the suboptimal fading factor when the existing process model is uncertain. The results of both numerical simulations and the field tests with GNSS/INS integrated navigation system demonstrate the effectiveness of the proposed algorithm.