Optimizing Slow Tool Servo Machining of Microlens Arrays Through Data-Driven Tracking Error Prediction and Trajectory Replanning

The application of microlens arrays (MLAs) based on slow tool servo (STS) machining is increasingly prevalent in the field of complex surface optics. However, current STS technique poses a significant challenge in balancing machining efficiency with surface profile accuracy, primarily due to the rapid variations in the spatial frequency of the microlens. To handle this dilemma, this study proposes a data driven tracking error prediction method and trajectory replanning approach. The specific method involves establishing a functional model that relates axis tracking error to the acceleration of the motion trajectory in STS machining. To further improve the predictive accuracy of this model, multiple calibration experiments were conducted, and a set of predictive model parameters was obtained based on least squares method. Furthermore, the original machining trajectory's acceleration was replanned, and combined with the predictive model results, to minimize the tracking error, thus optimizing the surface quality of microlenses in STS machining. To validate the effectiveness of the proposed method, diamond turning experiments were conducted on a three axis ultra precision machine tool. The experimental results showed that this method not only significantly improved the surface accuracy of the MLAs but also slightly increased the machining speed, resolving the predicament of STS machining MLAs.

Keywords: Microlens array, Ultraprecision diamond turning, Tracking error prediction, Trajectory replanning.

Download PDF