Real-Time Local Optimal Bezier Corner Smoothing for CNC Machine Tools

A straight line or G01 is the most common path in computer numerical control (CNC) contour machining. However, non-smooth corners introduce discontinuous movement along each axis during machining. The commonly used strategy is to redesign a corner into a smooth corner path. Therefore, optimal corner smoothing should simultaneously find an optimal corner path and a corner speed that satisfied the axis motion limits. To achieve this goal, we use particle swarm optimization (PSO) to find the optimal Bezier curve and the optimal corner speed, then convert the axis constraints into polynomial inequalities and use fitness selection to deal with axis acceleration and jerk limits. For real-time operation, we establish a look-up table with the optimal corner speed and corner path models for use by a CNC controller. In the experiment, the YCM NXV560A vertical machining center was used to conduct a real machining test on a Taiwan contour, and the feasibility of the controller to generate the optimal corner strategy for various linear paths was verified.

Automated summary

The study utilizes particle swarm optimization to find the optimal corner path and speed, and converts axis constraints into polynomial inequalities for handling. A look-up table is established for real-time operation.