Control of microassembly-robots by using fuzzy-logic and neural networks

Abstract

An automated micromanipulation desktop-station that has been developed at the University of Karlsruhe includes several piezoelectric microrobots capable of moving over long distances and manipulating in the range of a few nanometers. In this paper, the control problems of the microrobots and the sensoric problems in the micromanipulation station are discussed. The ability of an intelligent microsystem to adapt itself to the process requirements is, especially for assembly robots, very important. The microrobots should be able to operate in a partially defined environment and to guarantee a reasonable behavior in unpredicted situations. For this, the control methods are needed, which do not require an exact system model and which allow a reasonable compromise between the real-time information processing and the amount of input data; they must also make definite decisions based on vague or incomplete information. Two promising methods are fuzzy control and control by a neural network. The paper presents some approaches to improve the robot behaviour by using these advanced control methods.