Detecting Slowly Accumulating Faults Using a Bank of Cumulative Innovations Monitors in Kalman Filters

Abstract

Extended Kalman filters (EKFs) that monitor innovations over time have been demonstrated to be effective at detecting slowly accumulating measurement faults (Quartararo & Langel, 2020; Tanil et al., 2018). This paper first demonstrates that a single cumulative monitor becomes increasingly sensitive to measurement error model uncertainty as the accumulation interval increases, leading to false alarm and detection rates that can differ significantly from predefined design parameters. In response, a bank of finite-length cumulative innovations monitors is explored for fault detection in multisensor navigation systems. A novel extension to traditional covariance analysis (Covariance Analysis Including Expected Values or CAIEV) is developed to accommodate measurement faults and is used in addition to Monte Carlo simulations to present detection results for a variety of GPS fault profiles and inertial measurement unit (IMU) grades. Data for time-to-detect is presented alongside the position-domain bias induced by the fault at the time of detection. We show that the monitor bank can reliably detect the presence of faulty measurements after the position-domain bias has reached only tens of meters using tactical and aviation-grade IMUs for the cases considered, an improvement over other innovations-based techniques.