Use of fuzzy sets for systems modelizing and control: application to the speed regulation of a strongly perturbed motor

The theory of fuzzy sets has for some years now permitted a better grasp of certain systems in various different fields, notably, in that of automatic process control. This type of automatic control may be envisaged for a class of systems which are difficult to modelize and identify, but which may be run nevertheless by a human operator after a learning phase. Fuzzy set theory will then allow one to use the “imprecise” description a human operator will give of the control strategy he applies to his process, or of the information he can get from the system itself. This article presents the principle for elaborating a model based on the description of evaluating a process given by a human operator. An example of utilization of this model is then proposed, for insertion in a computer, in real time, of the speed regulation of a motor subjected to a variable perturbing torque; the model is represented by a fuzzy five dimensional relation constructed by means of a simplified fuzzy language translator. The control algorithm resulting from this model occupies 2 Kwords of computer memory and is executed in 30 ms, with results comparable to those of classical controls.