Publications

We consider mechanical systems, which are not known precisely. In detail we assume that the system parameters are unknown (lack of precise knowledge), moreover, we allow for uncertainties with respect to input and output, and to perturbations from the environment (ground excitations) as well. The consideration of such systems leads to the use of adaptive control. The aim is to design universal adaptive controllers, which achieve a pre-specified control objective. Almost all controllers existing in the literature offer the same drawback: (possibly unboundedly) monotonically increasing gain parameter. With respect to limited resources in applications it is necessary to design improved adaptation laws. We present simulations of several gain parameter models primarily in application to bio-inspired sensors.

We analyse two micro robots for 2-dimensional locomotion on a flat surface. Forced bending vibrations of continua are used by both systems which are excited by piezoelectric bending actuators. These vibrations are transformed by non classical legs to complex trajectories at the contact points between robot and environment. The locomotion direction of the system can be controlled by the excitation frequencies of the actuation element. Models are developed and investigated to describe important motion effects of the robots. Furthermore some experimental results are presented.

The motion of a finite chain of identical bodies along a straight line in a resistive medium is studied. The motion is excited and controlled by changing the distances between the bodies of the chain. This model simulates limbless animals (for example, earthworms) that move due to redistribution of mass along their bodies. The longitudinal motion of such animals can be provided due to a wave of deformation running through the animal's body forward, from the head to the tail, or backward, from the tail to the head. This principle of motion can be used for some kind of mobile robots. An advantage of these robots against conventional mobile machines is that the worm-like robots do not need propelling devices such as wheels, legs or caterpillars, can be readily made hermetic, and do not involve protruding components, which enables them to be used for motion in "vulnerable" media. Such robots can be used for motion in narrow slots and tubes to produce various technical operations in these media. Mobile systems able to move in resistive media without special propelling devices due to the change in their configuration or the motion of internal masses were studied, for example, in [1]-[4].

Motivatd by the technical progress, electromagnetic action principles are more and more used in micro systems. It is well known that the discrete positioning of magnetic beads like ferrofluid droplets can be realised using an array of electromagnets whereat the position of the beads depends on the arrangment of the electromagnets. This paper refers to the analogous positioning of magnetic beads independent from the arrangement of the electromagnets. Therefore a control strategy was designed to position magnetic beads like ferrofluid droplets, ferrogels or ferroelastomers. The derived control law allows the definition of dynamic properties like time and damping constant. To achieve these properties, the controller is adjusted based on a model of the plan. This model regards the dynamical movement of the mentioned materials in electromagnetic fields and is characerised by simpliefied, nonlinear, ordinary differential equations. The theoretical results are verified by means of a basic assembly allowing the positioning of spherical ferroelastomers on macro scale. The measurments illustrate that the theroretically derived position controller meets the requirements of a predefined time and damping constant.