Data-Driven Adaptive Vibration Noise Suppression Method for Electromagnetic Scanning Micromirror

For vibrational shocks and other uncertain noises such as measurement noise in electromagnetic scanning micromirror (ESM), a novel data-driven adaptive vibration noise suppression method (DAVNSM) is proposed in this article. In the DAVNSM, to achieve adaptive and robust filtering in the ESM, the noise is divided into Gaussian mixture distribution noise produced by multisource fusion and strong interference noise produced by vibrational shocks. The adaptability of the proposed filter is enhanced in complex noise environments by the Gaussian mixture modeland interactive multiple model. The robustness of the filtering process is further strengthened by a log-cosh robust loss function and the Levenberg–Marquardt iterative algorithm. Based on the generalized gradient technique, the corresponding Jacobi linearized form of the hysteresis Hammerstein model for ESM is obtained. Theoretical analysis of the convergence of the proposed DAVNSM is given through contraction theory. Moreover, experimental results demonstrate the validity of the DAVNSM.