HMI Risk without Over-Conservative Operation for Advanced RAIM Fault Detection and Exclusion

Advanced receiver autonomous integrity monitoring (ARAIM) significantly enhances the capabilities of traditional RAIM services using dual-frequency and multi-constellation satellite signals. Under high continuity requirements, fault detection and exclusion (FDE) functionality is imperative for ARAIM. FDE operates in two modes: performance prediction and real-time operation, and both of them need to calculate hazardously misleading information (HMI) risk as an integrity measure. However, FDE HMI risk becomes more complex compared to the well-established fault detection (FD) HMI risk due to the increased risk introduced by potential wrong exclusions. The existing predicted FDE HMI risk employs a probability accumulation approach, accumulating all possible exclusion-related HMI risks. Similarly, real-time HMI risk needs to consider multiple exclusion options for potential wrong exclusions. This paper highlights the possibility of excessive conservatism in the existing two FDE HMI risks. We propose a reexamination of the HMI risk definitions in the form of conditional probabilities for both modes. For predicted FDE HMI risk, it tightly selects the maximum risk among the exclusion options, rather than summing all of them. Regarding real-time FDE HMI risk, it is a posterior value of the excluded fault given the position solution and is independent of other unexecuted exclusion options. Based on these definitions, we derive new upper bounds for HMI risk in both modes, treating FDE as a whole, not directly reusing mature FD results. These bounds avoid over-conservative operation and already encompass the risk of possible exclusions. The simulation indicates that this paper's proposed FDE HMI risk has clear advantages in terms of PL tightness and algorithm simplicity.