Enis Ersü
Darmstadt University of Applied Sciences
Network
Latest external collaboration on country level. Dive into details by clicking on the dots.
Publication
Featured researches published by Enis Ersü.
IFAC Proceedings Volumes | 1984
Enis Ersü; Henning Tolle
Abstract The paper explains an unconventional learning control method based on assumptions in the liturature about human problem solving and information storage in neuronal networks. The on-line learning comprises two stages: The dynamic input-output behaviour of the process to be controlled is stored stepwise in a neuron-like manner into an associative memory as a predictive process model, the control strategy planned via this model by optimization of a goal oriented performance index is then trained in the same way into a second associative memory. As a general mapping the learned behaviour is in both cases in general nonlinear, and by this such a control design is especially suited for strongly nonlinear processes. Simulation demonstrate the applicability of the new control concept.
International Journal of Control | 1992
Henning Tolle; P. C. Parks; Enis Ersü; Michael Hormel; Jürgen Militzer
The paper discusses the importance of interpolating memories for different lay-outs of learning control loops, design considerations for such memories and the convergence of the learning process for one such memory. Finally, applications of learning control loops and future research in this field are considered.
IFAC Proceedings Volumes | 1985
Enis Ersü; K. Rathgeber; Michael Schnell; Werner Neddermeyer
Abstract The hardware implementation of a cartesian closed-loop control scheme will be presented which allows to define the dynamic behaviour of each degree of freedom of the cartesian coordinate system in a prescribed sense. The control system at joint level is designed by multivariate design methods with an additional feedforward component using the concept of inverse dynamics. To achieve high accuracy for cartesian motions quasi-continuous control mode with cartesian sampling periods of not greater than 5 ms is aimed at. A special purpose processor for calculation of kinematic and dynamic terns is designed and integrated into a multiprocessor architecture. This implementation concept with Robot Arithmetic Processor provides the necessary computational power and allows real-time cartesian closed-loop control which is also essential for cartesian sensory control tasks.
IFAC Proceedings Volumes | 1988
Henning Tolle; Enis Ersü
Abstract The growing amount of new, additional topics to be handled without extention of the time students stay at the university and without an extension of the number of available lecturers poses a difficult problem for the teaching body. The paper presents experiences gained with a seminar as a possible solution in the area of robotics. After a discussion of the general scheme of education for electrical engineering at the Technische Hochschule Darmstadt, the structure of the seminar and experiences gained with it are described. The paper concludes with some remarks on another solution chosen with respect to introduce artificial intelligence approaches to control engineers.
IFAC Proceedings Volumes | 1987
Enis Ersü; S. Wienand
Abstract The paper discusses the real-time implementation of an associative-memory-based learning control scheme with PI-controllers for nonlinear processes. Starting with a pre-assumed PI-controller which only has to stabilize the process the controller parameters are optimized on-line by a predictive optimization. This optimization uses for prediction the model of the process stored in an associative memory which is also learned on-line. The situation-dependent optimized controller parameters are also stored in an associative memory. The concept is a modification of the LERNAS-system (Ersu,1984), which is also shortly described and compared to the system described here. Some experimental results with a nonlinear pH-control demonstrate the performance of the system.
IFAC Proceedings Volumes | 1987
Enis Ersü; Gunther Kegel
Abstract Optical proximity sensors are low-level vision sensors delivering low- and mid-range multidimensional information about the robots end-effector environment. Due to the fast signal processing they can be easily integrated in real-time robot control tasks. The paper presents the basic mechanical, hardware and software design principles of such a sensor, which uses distance measurement via optical triangulation as the basic method. For special robot tasks, special mechanical and hardware arrangements of the basic sensor type are needed. Two examples are shown for demonstration purposes. Possible applications are simple distance sensor devices, two-dimensional orientation sensors and optical robot teach-in units. Accuracy and efficiency of the sensor system are documented by using the sensor for recognizing holes and following arbitrary unknown contours.
Archive | 1988
Enis Ersü; Henning Tolle
neural information processing systems | 1987
Enis Ersü; Henning Tolle
Archive | 1992
Jan Matthiesen; Henning Tolle; S. Wienand; Enis Ersü
Archive | 1983
Enis Ersü; Werner Neddermeyer; Volker Tews