Robust Supervisory Control for Cyber-Physical Systems based on Discrete Timed Petri nets

The widely available sensor data from IoT enables the implementation of real-time controllers in factories to efficiently track the progress of production activities and status of resources in the systems. However, failure prone resources or machines in factories frequently lead to negative effects on the performance of production systems. In addition, due to contention of resources involved in different production processes, the Cyber-Physical Systems (CPS) may be brought to undesirable states in which all or parts of the production activities are blocked, in circular waiting or starvation. Recently, there are relevant studies on the impact of resource failures on Cyber-Physical Systems on a discrete timed Petri net (DTPN) model. However, the capabilities to avoid the undesirable states in CPS based on discrete timed Petri nets models are not explored. This paper aims to study the properties of a discrete timed Petri net supervisory control method that can avoid undesirable states while achieving the goal of production by a specified time constraint. These capabilities to avoid the undesirable states in CPS based on DTPN models are characterized by a property. The proposed supervisory control method is illustrated by an example.