A BPNN-PID based long-stroke nanopositioning control scheme driven by ultrasonic motor

Abstract

In this paper an innovative nanopositioning control scheme for different travel lengths is proposed. A commercial ultrasonic motor HR4 and its driver AB2 are employed to generate 3-mode motions (AC, Gate, and DC modes) to accommodate different travels, speeds and resolutions. For precise displacement feedback, a new displacement sensor LDGI (linear diffraction grating interferometer) is developed to meet the requirements of both long range and nano resolution. A key technology in this study is the proposed positioning control algorithm for the linear stage driven by HR4 and AB2. A 3-mode digital PID controller with a self-tuning module by back propagation neural network (BPNN) is developed for multi-scale and all-in-one motion control of 3 modes. Both experiments and software simulation show that this software-based controller developed by LabVIEW has good capability to overcome the uneven friction of the sliding plane and to lock the final position stably. The highlight of this 3-step motion control system is first to drive the table by AC mode at a low and stable speed in millimeter per second scale, then to move close to the target point by Gate mode with the positioning error less than 100 nm in micrometer per second scaled speed, and finally to adjust and hold at the target point by DC mode in nanometer per second scaled speed. In the experiments of different travels up to 15 mm, calibrated by a commercial laser interferometer, the positioning accuracy is proved within 10 nm with standard deviation less than 5 nm and the final position locking can be limited to 3 nm.