To address the challenge of simultaneously achieving high-precision trajectory tracking and active vibration suppression in precision motion stages under complex dynamic operating conditions, this paper proposes a deep learning-based multisensor fusion control method. By exploiting the complementary measurement characteristics of position and acceleration sensors, a multisensor fusion model is constructed for dynamic state estimation, allowing control-relevant features related to motion states, lumped disturbances, and vibration evolution to be extracted from multisource temporal data. Based on a baseline feedback controller, a fusion-estimation-driven compensation strategy is then designed to incorporate disturbance compensation and vibration suppression into the closed-loop framework, thereby improving tracking accuracy and dynamic smoothness. Experiments, including standard trajectory tracking, disturbance rejection, and high-speed dynamic tests, are conducted on a single-axis precision motion stage and used to compare the proposed method with a baseline feedback controller and a conventional fusion/observer-enhanced method. Results show that the proposed method achieves lower tracking errors, smaller vibration peaks, faster residual vibration decay, and better disturbance tolerance than the comparison methods. These findings indicate that the proposed fusion-control framework is effective for improving the overall performance of precision motion systems.

precision motion stage, multisensor fusion, active vibration suppression, trajectory tracking