Asymmetric Dual Redundant Inertial Sensors: A Method for Fault Detection and Diagnosis

This article proposes fault detection and diagnosis (FDD) algorithms for an asymmetric redundant inertial measurement unit (ARIMU), which consists of both high-precision and low-precision sensors. The algorithm combines the generalized likelihood test (GLT) and the fading sequential probability ratio test (FSPRT) using centered parity vectors. It addresses the challenge of identifying faults in an orthogonal dual-structured redundant system with varying device precision by presenting a deviation-consistency-test fault diagnosis method. Numerical simulations and semi-physical experiments confirm the effectiveness of these algorithms. Compared to conventional SPRT and GLT algorithms, the proposed combined algorithm enhances fault detection speed by 70.07% and 87.73%, respectively. Moreover, it can track multiple faults and accurately locate faulty inertial devices using the deviation-consistency-test diagnostic method. This work greatly improves the sustainability and availability of ARIMU system, offering an innovative engineering solution to boost the reliability of inertial measurement unit (IMU).