A fault detection method for electromagnetic scanning micro-mirror

Abstract The electromagnetic scanning micro-mirror (ESM) exhibits pronounced hysteresis, characterized by its non-smooth, rate-dependent, and multi-valued mapping properties. These inherent characteristics pose significant challenges for precise state estimation and reliable fault detection. For the fault detection of ESM affected by unknown input and white noise, a fault detection method is proposed. This method integrates disturbance decoupling via an unknown input observer (UIO) with white noise suppression through the propagation of state estimation error covariance. In the proposed method, the gain of UIO is derived by minimizing the state estimation error covariance, which is defined as the loss function with the objective of reducing the influence of white noise on the state estimation. Furthermore, progressive processing of unknown inputs and white noise enhances state estimation robustness in the proposed method. In addition, a fault detection threshold is designed based on the idea of statistical characteristics, and the fault detection task for ESM is accomplished by a threshold comparison method. Finally, the effectiveness of the designed fault detection method is verified through simulation and experimental results.