Optimization of the STEP-NC compliant online toolpath generation for T-spline surfaces using convolutional neural network and random forest classifier

Deep implementation of principals of the intelligent STEP-NC compliant manufacturing implies enabling a CNC system with integrated CAM functionality to perform all tasks autonomously. This means that most of the responsibilities of a CAM engineer need to be delegated to the CNC system. When it comes to manufacturing of freeform surfaces, the implementation of this approach becomes very challenging. One of the most important issues is the autonomous online toolpath generation. Making a choice of optimal machining strategies and other manufacturing parameters, that might be a trivial task for an expert, turns out to be not an easy problem for a machine, and often can be hardly solved using the traditional approach of explicit programming. Therefore, this paper proposes a method to optimize toolpath generation for T-spline surfaces, in particular, the process of choosing an optimal machining strategy for a given surface region using Machine Learning. The selection model has been trained to make a choice of the appropriate freeform machining strategy (two different strategies have been considered) based on the shape of an input surface.