Real-Time Simulation of Robot Controlled Belt Grinding Processes of Sculptured Surfaces
Industrial robots are introduced to belt grinding processes of free-formed surface with elastic wheel nowadays in order to obtain high quality product and high efficiency. However, it is a laborious task to plan grinding paths and write programs for the robot. To release people from it partially, it is necessary to simulate the belt grinding processes which are useful for path generating and dynamic robot control. In this paper, we present a framework of the robot controlled belt grinding simulation system and some key issues in it. We enhance the global removal model to local process model, which can simulate the grinding process more exactly. We also point out the bottleneck of the real-time simulation and put forward a neural network based regression method to meet this difficulty. At the end of the paper, some simple simulation examples are given.