H. Brugger

A Rapid Method for the Preparation of a One Dimensional Sequence-Overlapping Oligonucleotide Library

Abstract A new synthesis modul was constructed to prepare one-dimensional libraries of sequentally overlapping oligonucleotides on surface modified polypropylene tapes. The use of such a library for the diagnostic screening of a point mutation in the regulatory protein gene p53 is described.

Optimal feedback position control for an autonomous mobile robot

Wheeled autonomous mobile robots underlie nonholonomic constraints and therefore they belong to a special class of nonholonomic systems. For this class of systems feedback stabilization plays a central role because they cannot be stabilized by continuously differentiable time invariant control laws. In this paper a time varying feedback control law that minimizes the input energy is designed for solving the optimal feedback control problem for high precision positioning.

Motion control module for an autonomous mobile robot

We analyse the motion control module implemented on the autonomous mobile robot B21 as part of the specific software architecture developed for the SFB 527, called integration of symbolic and subsymbolic information processing in adaptive sensory-motor systems. The motion control module is first located the lowest layer of the architecture. Its functionalities are path control skill, position control skill, via-point navigation skill, and put-through skill. The interfaces to other modules and the configuration states are defined. The communication tasks are implemented independently of the functionalities in separate threads using the SMART/sup Ulm/ software frame.

Optimal Control for a Synchronous Driven Unicycle-Like Autonomous Mobile Robot

In this contribution the autonomous mobile robot B21 manufactured by Real World Interface is modelled as a synchronous driven system leading to a nonholonomic mechanical system described by three coupled nonlinear differential equations. The obtained system is completely controllable. A preliminary change of coordinates together with a change of inputs transforms the original system into one in chained form. Now it is possible to approach the design of an optimal controller. The input vector is to be chosen to drive the system from any initial to any final state configuration while minimizing the input energy. A quadratic performance index is defined integrating over the square norm of the transformed inputs. The Hamiltonian associated with this optimal control problem is calculated with the co-state vector. From the Maximum Principle the optimal control law is found by minimizing the Hamiltonian with respect to the transformed inputs together with the equations for the co-states. The obtained system is solved analytically and the optimal input vector is also calculated analytically to obtain the optimal control law for the defined performance index. For obtaining experimental results the designed control law is implemented on the robot B21. A real time software package has been developed to comfortably implement the designed controller on the robot. Several experiments are carried out the results for one specific task, i.e. driving backwards and sidewards simultaneously, are shown. The experimental results are highly satisfactory because the driven paths are smooth and for the steering inputs smooth functions axe used.

An Automated Method to Create a One Dimensional Array of Chemically Synthesized Oligonucleotides

In this paper we describe a method to synthesize a variety of different oligonucleotides in a quasi parallel procedure on the surface of a polypropylene tape. The relation between necessary chemical coupling steps n and thus generated oligonucleotides m with a length of e.g. 15 monomers is: m = n − 14. To get a high synthesis throughput and the possibility of an overnight run, the system has been fully automated. To start the synthesis, only the desired sequence programming is necessary without further control of the system.

Global asymptotisch stabile Positionsregelung für einen autonomen mobilen Roboter

Grundvoraussetzung fur einen autonomen mobilen Roboter sind zuverlassige Regelalgorithmen fur die Bewegungsfuhrung. In diesem Beitrag wird uber den Entwurf von Positionsreglern und deren Implementierung auf einem autonomen mobilen Roboter berichtet. Durch die Verwendung des Invariance Principle von La Salle und Lefschetz wird der Beweis der globalen asymptotischen Stabilitat des geregelten Systems mit Hilfe einer Ljapimow-Funktion wesentlich vereinfacht. Beim Reglerentwurf wird ein naturliches Fahrverhalten angestrebt, wobei die Begrenzungen der Stellgrosen eingehalten werden sollen. Die zur Implementierung auf dem Roboter eingesetzten Softwarehilfsmittel werden erlautert. Die hohe Gute des Reglers wird mit experimentellen Ergebnissen nachgewiesen.