Position tracking performance enhancement of linear ultrasonic motor using iterative learning control

Abstract

The highly nonlinear characteristics of the ultrasonic motors have posed problems for its high precision position control. There are two ways to tackle the nonlinearities in the system: (a) make use of advanced nonlinear position control schemes to compensate for the nonlinearities as a whole or (b) compensate the nonlinearities partially by providing a suitable feed-forward signal and then make use of a linear feedback controller for the outer loop position control. In this paper, nonlinearity introduced due to deadzone is first compensated using a feed-forward signal and position control of the linear ultrasonic motor (LUSM) is carried out using a conventional PI position controller. The performance of such a closed loop control system is evaluated experimentally. The performance comparable to those that can be obtained using computationally intensive non-linear control schemes proposed in V. Utkin et al., D. Q. Zhang and S. K. Panda (1999) can be obtained using linear controllers. Though PI position controller exhibits very good set-point tracking, its performance deteriorates in case of the load variation or for tracking time varying references. In order to improve the tracking performance a robust sliding mode controller is then implemented. Experimental results obtained validate the improved tracking performance in comparison with PI position controller. For periodic reference position tracking the sliding mode controller exhibits periodic tracking error profile. For further improving the tracking performance an iterative learning control is augmented to the sliding mode controller. Experimental test results obtained show that the periodic tracking error can be suppressed by a factor of ten using the proposed iterative learning control scheme.