Novel analysis of Petri-net-based controllers by means of TCT implementation tool of supervisory control theory

Abstract

The control of discrete event systems (DES) has been widely studied in the past two decades. Finite-state automata (FSA) and Petri nets (PN) are the two principal modelling formalisms for this study. Supervisory control theory (SCT), based on language and FSA concepts, is a well established framework for the study of discrete event control systems (DECS). PN-based approaches to the control design have been considered as an alternative framework. In the PN-based control of DES, given an uncontrolled PN model of a system and a set of specifications, a PN-based controller consisting of monitors (control places) is synthesised to solve the problem. In general, forbidden-state specifications are considered. Another heavily studied specification is to obtain the live system behaviour (non-blockingness in SCT terminology) for a given PN model by computing a PN-based controller. Unfortunately, PN-based analysis tools cannot deal with uncontrollable transitions. Therefore, to date there is no general technique for the correctness analysis of the computed PN-based controllers. This paper proposes a novel and general methodology to carry out the correctness analysis for the computed PN-based controllers by using the TCT implementation tool of SCT. Three examples are considered for illustration.