Modellingand supervisory control of hybrid dynamical systems via fuzzy l-complete approximation approach

This paper proposes a fuzzy l-complete approximation technique based on Willems' behavioural systems theory for a class of hybrid system with symbolic inputs and quantized measurements. The discrete controller (or supervisor) is designed based on the abstracted (or approximated) system and high-level specifications. Furthermore, we show that the set of external behaviour (i.e., state set) of the fuzzy l-complete approximation based on the notion of 'guardian states' contains or at least equal to the behaviour set generated by traditional l-complete discrete abstraction. Finally, several simulation examples are presented to show the effectiveness of the proposed method for supervisory control of a special class of hybrid systems.