Immune System Based RBF Neural Network Modeling for Machine Tool Thermal Error

An artificial immune RBF neural network(AIRBF) was presented after comparing human immune system with artificial RBF neural network.This modeling method has the ability of self-organization and on-line studying.It overcomes the disadvantage of the existing models that they could not change their model structures and real-time learn new information.An experiment on a CNC lathe was conducted to verify AIRBF and a comparison with traditional RBF neural network was made.The result indicates that the structure of AIRBF is simpler and AIRBF can accommodate the nonlinear nature of thermal elastic process.Thus,AIRBF can get better prediction accuracy.