Inverse Adaptive Control
Principal Investigator: István Kollár

Control of various industrial processes is a topic of intensive research world-wide. Since the spread of the very successful PID controllers, so-called modern control methods have also been introduced for robust control, mostly based on H_inf design methodologies. However, there is a significant gap between industrial practice and academic research. Sophisticated methods did not gain widely used applications, since practical people often prefer simple concepts and easy-to-implement controllers to sophisticated principles. Therefore, it is very important to develop methods which can be easily understood and accepted in the industry.

The new book of Bernard Widrow and Eugene Walach, "Adaptive Inverse Control" (Prentice Hall, 1996) presents such a concept. Instead of feedback, it suggests a controller in series with the plant, realizing the inverse of its transfer function. The feedback is also present, but rather in the adaptation process than directly in the control. By proper adaptation, the inverse controller strives to minimize the error at the plant output, much like in conventional controllers. Having the new concept, it should be brought to usage. This means implementing all details by available programs. The main program Matlab offers an ideal environment for this.

The purpose of this project is to develop a new toolbox on Inverse Adaptive Control. This should be offered not only to practicing engineers, but also to education. The aim is to realize all the concepts of the above book in Matlab, investigate their capabilities on practical examples, explore the usability of resonator-based adaptive filters developed at the DMIE, and ultimately, realize an easy to use Matlab toolbox with application examples. The work is done in cooperation with the Information Systems Laboratory (ISL) of Stanford University (CA, USA).


Cooperating partner at ISL: Bernard Widrow.