Neural Network-Based Error Control in CNC Machining of Mechanical Components
This paper describes in detail a robot milling machining error compensation method based on binocular vision system and neural network-based.The PowerVision1500D binocular vision instrument is used to achieve accurate measurement of tool end position error in robotic milling machining.Through the coded marker points configured on the workpiece and robot spindle,the binocular vision system is able to acquire and calculate the actual position of the tool TCP under the workpiece coordinate system in real time.Further,a prediction model was developed using a particle swarm optimised back-propagation neural network for simulating and predicting the relationship between the robot's positional attitude and the positional error of the tool TCP,so as to plan and adjust the milling trajectory in advance.Finally,the integrated error compensation method is experimentally validated on an aerospace aluminium 7075 material component,and the results show that the compensation strategy significantly improves the machining accuracy and efficiency.
NETL
NSTL
万方数据
