Elementary Siphon-Based Robust Control for Automated Manufacturing Systems With Multiple Unreliable Resources

• Published in: IEEE access Vol. 7; pp. 21006 - 21019
• Main Authors: Li, Xiaoyan, Liu, Gaiyun, Li, Zhiwu, Wu, Naiqi, Barkaoui, Kamel
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Automated manufacturing system; Automatic control; Automation; Computer Science; Control systems; deadlock prevention; Failure; Failure analysis; Law; Modeling and Simulation; Monitoring; Monitors; Petri net; Recovery; Robust control; Robustness; Siphons; System recovery
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2019.2897753

Resources in a real-world automated manufacturing system may be unreliable, thus paralyzing many existing deadlock control strategies. This paper defines robust legal markings and proposes a two-step robust deadlock control policy for systems of the simple sequential process with resources based on elementary siphons. The first step, called elementary siphons control, derives a live controlled system without considering resource failures. The second step deals with failure-induced deadlocks control issues. With resource failures taken into account, recovery subnets modeling resource failures and recoveries are added to the first-step-derived system, which may cause the occurrence of failure-induced deadlocks, resulting in an unreliable controlled system. In order to make the system reliable, the definition that a resource failure is observed by a monitor is developed. A scenario is exposed that a resource failure does not need to be observed while can be observed by the added monitors. Arcs connecting recovery places and transitions of the initial system are added based on the obtained scenario. The concept of the extend the adjoint set of a siphon is presented due to a resource failure that needs to be observed while cannot be observed by the added monitors, leading to the modification of the arcs added in the first step. Consequently, a robust controlled system is acquired. Finally, examples are presented to demonstrate the proposed method.