Self-organizing qualitative multimodel control

Abstract

Qualitative control is a promising technique for controlling dynamic systems which are ill-defined or too complex to be modelled analytically. In this paper, a Self Organizing Qualitative Control (SOQC) is proposed for controlling dynamic systems with incomplete knowledge on their parameters. This technique is based on a learning approach which uses a qualitative multimodel representation to reduce parametric uncertainty and synthesize a fine control action. The standard problems“benchmark examples” have been selected to apply the proposed control method. To show the potential of this method for controlling interconnected systems, the control of a two coupled tanks system is achieved using the same control technique. The simulation results are satisfactory for achieving tracking and regulation objectives.