Integrity Monitoring Algorithms using Filtering Approaches for Higher Navigation Performance: Consideration of the non-Gaussian GNSS Measurements

摘要

For safety-critical applications of global navigation satellite systems (GNSS), such as aviation and missile navigation systems, it is important to detect and exclude faults that cause accuracy and integrity risks, so that the navigation system can operate continuously without performance degradation. For high accurate systems, the integrity monitoring function needs to detect and exclude small biases. And as more independent GNSSs (ex. GALILEO and GLONASS) are being available, we need to take into account simultaneous faults of multiple satellites. It also estimates protection level that determines availability of the navigation system. With conventional snapshot RAIM algorithms, it is difficult to detect small errors and simultaneous multiple faults. Assumed that we know the system dynamics, filtering algorithms, such as the Kalman filter, can provide better monitoring performance than the snapshot algorithms can, because the filter reduces noise level of measurements. However, because the Kalman filter presumes that measurement noise and disturbance follow the Gaussian distribution, its performance might degrades if the assumption is not right. To address this problem, we propose a fault detection and exclusion algorithm using particle filters. The particle filters are popular filtering methods to estimate states of a general dynamic system. It can deal with any system nonlinearities or any noise distributions using sequential Monte Carlo method, and present the posterior distributions of the states completely. Because GNSS measurement noise does not follow the Gaussian distribution perfectly, the particle filter can estimate the posterior distribution more accurately; therefore it has better integrity monitoring performance. Also, if the system has high non-linearity, the performance is getting better. Additionally, an integrity monitoring algorithm using the Gaussian Sum Filter is proposed as an alternative to the particle filter method that needs high computational load. The paper describes the detailed algorithms, and shows simulation and experiment results to evaluate integrity monitoring performance of the algorithms. The proposed algorithms detect 20% smaller faults and generate 30% lower protection levels than the conventional filtering methods that use the overbounded sigmas considering the non-Gaussian measurement distribution can. The results show that the proposed algorithms that do not use the overbounding method can provide better accuracy and availability performance just by changing the filtering algorithm.