for
more information download a tared PDF-version
(900 kB) or send me a mailProf. Dr.
G. Schweitzer, Examiner
http://dmtwww.epfl.ch/imt/asl/people/siegwart.html
Prof. Dr.
R.Y. Siegwart, Co-examiner
stepping motion, inertial drive, impact drive, micropositioning, piezo-electric, positioning, micromotor, multi-degree-of freedom, nanorobot, micromanipulator
Micro- and nanotechnology is a key issue in today's and tomorrow's development of advanced products. New tools are and will be needed to automatically handle and assemble micro-sized structures with sub-micrometer precision, or to give human beings the capability of operating in these tiny dimensions. This work focuses on the mechanics of micromanipulation. After some basic considerations, the specifications for a high precision robot structure are defined. Among them, one of the most challenging problems is to position a robot extremely accurately (10 nm) within a large workspace (1 cm3), i.e., with a spatial dynamics of 10-6. The only way to achieve this accuracy is to use direct measurement of the relation between the target object and the microtool used to treat this object. For the mechanical setup, this strategy reduces the problem of accuracy to a problem of resolution. Special attention has to be paid to friction, more precisely, to the stick-slip phenomenon, as it is the main limiting factor for the positioning resolution. Stepping actuation principles are able to cover the above dynamic range, since they combine both a high resolution and a theoretically infinite workspace. Driven by piezoelectric actuators, these mechanisms are very simple in design and easy to control. Therefore, they are well suited for micropositioning tasks. Two kinds of stepping principles are investigated: The crawling principle and the inertial principle. Mathematical models are developed which explain the mechanisms behavior. Experiments are performed to verify these models. In a second part, we discuss in detail the combination of single actuators to multi-degree-of-freedom mechanisms. The aim is to provide high resolution motion within a large 3D working range. To avoid compliant, friction-limited designs, we propose parallel manipulators with flexible joints and links. Two different structures with three degrees of freedom each will be investigated: A planar inertial drive for motions on a horizontal tread and an Inchworm driven spatial structure to attain the complementary three degrees of freedom. In contrast to macro-robotics, the main restrictions to the kinematic configuration of micromanipulators originate from the sensor, i.e., the microscope used to supervise and control handling and assembly processes. Our approach is a multi-arm robot able to operate in the limited space under the microscope. All these developments were implemented and tested within the ETHZ NanoRobot System. Moreover, benchmark tests proved the feasibility. With simple microassembly tasks performed with teleoperation we demonstrated the ability for micromanipulation and future applications. The accuracy of the closed-loop system is evidenced by high precision motions guided by a feedback controller based on computer vision.
for
more information download a tared PDF-version
(900 kB) or send me a mail
