Iterative learning based trajectory generation for machine tool feed drive systems

In machine tool performance, a fundamental factor is an axial movement which is driven to track a desired trajectory. Not only tracking errors in each drive axis but also contour errors, which are directly related to the machined shape of a workpiece, should be considered. Although most existing contouring controllers are based on feedback control, this paper proposes an embedded iterative learning contouring controller (EILCC) by considering both tracking and contour errors. The proposed control iteratively modifies the reference trajectory of each drive axis to reduce the contour error. The proposed controller can be directly applied to commercial machines currently in use without any modification of their original controllers. The proposed method has been experimentally verified through a biaxial feed drive system on a sharp-corner trajectory which normally leads to a large contour error around the corner due to the discontinuity. Comparison with a conventional iterative learning contouring controller (CILCC) was done so as to evaluate its performance. Experimental results have shown that the contour error converges within a few iterations, and the maximum contour error can be reduced by about 49.2% as compared to the CILCC.