Real-time pose measurement for the cutter of a virtual axis machine tool based on a RBFNN

Abstract

Real-time pose measurement for the terminal tool of a virtual axis machine tool is still one of obstacles to achieve high-precision control and industrialization of virtual axis machine tool in the field of digital control machining. The pose measurement of the tool for a 6-DOF virtual axis machine tool is studied in this paper. Firstly, the kinematics analysis of the virtual axis machine tool is made, then the tool pose and its inverse kinematics results are used as the neural network training samples, and the RBF neural network with a self-adaptive structure is established, so that the mapping from a joint variable space to a work variable space is realized for the virtual axis machine tool. Finally, the real-time pose measurement of the tool is achieved by using the trained neural network and the motion states of the active joints which can be detected easily. Experimental results show that the method of measuring the pose of the tool of the virtual axis machine tool based on the RBF neural network with the self-adaptive structure is not only of feasibility but also of high-precision, which establishes the basis for direct closed control of virtual axis machine tool.