Simultaneous Optimization of the External Loop Parameters in an Adaptive Control Based on the Co-operation of Uniform Procedures
József K. Tar*, Imre J. Rudas*, Ladislav Madarász** and János F. Bitó***
*John von Neumann Faculty of Informatics Institute of Mathematical and Computational Sciences, Budapest Polytechnic, H-1081 Budapest, Népszínház utca 8.Hungary
**Department of Cybernetics and Artificial Intelligence, Technical University of Kosice, 042 00 Kosice, Letná9/b, Slovak Republic
***Centre of Robotics and Automation (CRA), John von Neumann Faculty of Informatics Institute of Mathematical and Computational Sciences, Budapest Polytechnic, H-1081 Budapest, Nápszínház utca 8.Hungary
Control approaches based on fuzzy controllers, artificial neural networks that is on “soft computing” (SC) are generally constrained by standardized formal procedures and structural restrictions. In general these restrictions guarantee the solution for a quite wide class of prolems. However, “orthodox” application of SC may result in overcomplicated control not necessarily exploiting the peculiarities of the given task. This often manifests itself in a huge number of unknown parameters to be identified or tuned. In this paper non-conventional integration of simple elements as classic PID/ST, regression analysis, saturated sigmoids, fuzzy sets and uniform structures obtained from the Lagrangian Mechanics is presented. The great majority of the parameters applied are either independent of the particular problem to be solved or have simple rigid tuning rules expressed in the language of the conventional fuzzy controllers. Due to the relatively simple structure of reduced number of the free parameters real time tuning can be carried out. In this paper the simultaneous optimization of the free parameters is considered in details. It is concluded that via this non-conventional integration simple and efficient adaptive control can be constructed for robots of very approximately known dynamic properties and being involved in unknown environmental dynamic interaction.
This article is published under a Creative Commons Attribution-NoDerivatives 4.0 International License.