Research on acquisition and feature analysis of GMA status signal based on grating sensors

摘要

The new GMA experimental platform based on grating sensors and modeling method with BP neural network are proposed. The various parameters of GMA in the working process, such as driving voltage, prestress, displacement, are gathered and the effects are analyzed, then BP neural network is used to build the GMA model according with the real features of driving voltage and prestress. The model is verified by experiment with actual data, it can predict the actuating output of GMA greatly, and the predict error is in the range of 0-0.004nm.