Improved fault detection and tolerance method based on local test and adaptively robust filtering for INS/GNSS integration

摘要

The abrupt fault contained in GNSS measurement will reduce the accuracy of SINS/GNSS loosely-coupled integrated navigation. To solve this problem, the performance of residual chi-square test method is analyzed, on this basis, an improved fault detection and tolerance method based on robust adaptive adjustment is proposed. The local test is implemented to detect and identify the faulty observation. Aiming at the problem that the small size of fault is hard to detect, the gain matrix is adjusted adaptively by using the robust adaptive factor to improve the fault tolerance ability for the missed detection. This in turn improves the detection performance of local test method for the small size of fault. The simulation results show that compared with conventional methods, this method has better performance for the abrupt fault especially small size of fault, and can improve the accuracy and reliability of integrated system effectively.