Položajno vodenje robotskega mehanizma s piezoelektričnimi aktuatorji : doktorska disertacija

slovenian

doktorska dela

2011

Univerza v Mariboru, Fakulteta za elektrotehniko računalništvo in informatiko

This thesis presents modelling, design and position control of a planar robot mechanism with piezoelectric actuators. Robot mechanism consists of a glass frame and two piezoelectric actuators. The length displacement of piezoelectric actuators is transmitted through a glass frame with built-in transfer ratio at the tip of the robotic mechanism, where it is represented as his working shift. The size of the robot mechanism is approximately 40x40 millimetres and the size of the working plane is approximately 25 by 25 micrometres. By employing piezoelectric actuators the tip of robotic mechanism can be positioned with sub-micrometre accuracy. However, due to the nonlinear hysteresis effect of piezoelectric actuators, this functionality is not guaranteed without the appropriate position control. Concerning these observations, a system for position control of robot mechanism has been designed and implemented. The implemented system includes voltage power amplifiers for piezoelectric actuators, the metering equipment for measuring the actual position of the robot mechanism's tip and control-processing unit with software equipment and algorithms for position control. Due to the nonlinearity of controlled plant, the use of traditional control methods, suchas PI controller, does not give very satisfactory results. Therefore, for evaluation and compensation of plant nonlinearities (hysteresis in piezoelectric actuators), we use feedforward neural networks, which are known as a universal tool for evaluation and compensation of nonlinear phenomena. Experimental results obtained on real robot mechanism confirm, that the use of neural networks can improve the performance of traditional open-loop and also close-loop control methods for position control of robot mechanism.

07.08.2023