Sensor Fusion for Active Vibration Isolation System with Double Noise: An Adaptive Kalman Filter Approach

This article presents an adaptive Kalman filter (AKF) with double noise for active vibration isolation (AVIS) system. Due to the time-varying vibration input and unknown sensor noise, it is difficult to obtain the process noise covariance matrix and the measurement noise covariance matrix. The imprecise noise covariance matrixes may result in Kalman filter (KF) divergence to further reduce the performance of the AVIS. Consequently, to achieve the time-varying process noise covariancematrix estimate, the internal forgetting factor of maximum likelihood estimation is reconstructed. The recursion covariance of the state space equation is used to estimate the measurement noise covariance matrix, and the innovative hypothesis is used to select the superior estimate of double noise covariance matrix. The convergence of the AKF is demonstrated. A single-degree-of-freedom AVIS is built to validate the effectiveness of the proposed method. Some experiments show that the vibration isolation performance is improved by 26 dB@30 Hz.