Precision control of a micromanipulator

• Main Author: Low, Kay Soon
• Other Authors: School of Electrical and Electronic Engineering
• Format: Research Report
• Published: 2008
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation::Control engineering
• Online Access: http://hdl.handle.net/10356/14177
• Institution: Nanyang Technological University

The startup grant SUG 36/04 provides the principal investigator a funding for conference support. As there is no funding for the equipment and manpower, the work reported in this report relies on some existing experimental system and manpower from other resources. With the advancement of technology, accurate positioning systems are increasingly required in various industries to perform difficult task and to improve the productivity so as to lower cost. In this study, we conducted our work on three experimental systems, namely 1. a harddisk drive servo manipulator, 2. a MEMS based variable optical attenuator, 3. a precision linear stage. For the harddisk drive manipulator, we formulate the control problem as a multiobjective optimization problem. While some of the objectives are constraints objectives, some of them are simply optimization objectives. We have developed a novel approach that is based on a new multiobjective genetic algorithm that can tune the controller parameters such that the resultant precision servo can perform and meet all the requirements. Unlike the traditional approach, our approach could place higher priority for the constraint objectives than the optimization objectives. Experimental results based on a commercial harddisk have verified the effectiveness of the proposed approach. For the MEMS based variable optical attenuator, the device is being used to control the power in the all optical network. An advanced control approach based on the modelpredictive control scheme has been developed. Both the electromechanical behaviour and the optical model have been adequately modeled. Through both the simulation studies and some experimental measurements, it has been shown that the proposed control method is superior than the existing method based on the PID controller. For the third system, a linear precision machine is used. It is a linear stage that has a positioning accuracy of 1 μm and a peak speed above 1 m/s. In this system, a permanent magnet DC linear motor has been used as the actuator to eliminate the needs of mechanical transmission from the rotary to linear motion. The key to perform the precision movement and achieving fast and accurate closed loop response is the development of a state space model predictive controller with a dynamic friction compensation scheme. Moreover, the machine has been optimally tuned using an on-line genetic algorithm. Further work on friction estimation has also been attempted using a pulse based neural network. In conclusion, the study has been successfully carried up to work on three different precision control systems. Advanced control approach and evolution techniques such as genetic algorithm and pulse base neural network have been developed. They have been applied to the practical systems and the results have shown significant performance improvement.