Adaptive observer-based fault detection and diagnosis for aircraft landing gear actuators

Abstract

This paper aims to present an observer-based diagnostic structure for actuator fault detection in the aircraft landing gear avionics system. The observer estimates system states by comparing measured and estimated gear positions, generating residuals that are used to detect and identify actuator faults. The method accounts for all movements of the nose and main landing gear. By modeling actuator faults as additive disturbances and adapting the residual analysis to operational phases, the method ensures high sensitivity to faults while preserving robustness, making it suitable for real-time aerospace applications. An inference step, acting as a decision block, determines the presence of faults and identifies anomalies from residual signals generated by the observer and sensors. This study emphasizes the value of methodical adaptation over theoretical novelty, demonstrating that a classical observer, when carefully tuned to the operational dynamics of landing gear actuators, can yield an effective, interpretable, and certifiable fault diagnosis solution for aerospace systems. Simulation results demonstrate accurate state estimation and effective fault detection under both nominal and faulty conditions, confirming the approach’s suitability for aerospace applications.