A Robust Control Approach to Automated Manufacturing Systems Combining Absorbing and Distributing Characteristics
Xiaojun Wang, Hesuan Hu · IEEE Transactions on Systems Man and Cybernetics Systems · 2019
The subject of deadlock control for automated manufacturing systems has been extensively studied in the past several years. For most of them, researchers have supposed that shared and dedicated resources never fail. As any manufacturing practitioner knows, resource failures, e.g., defective parts, faulty sensors, blurred signals, and broken actuators, are a common problem. Hence, the robust supervision and control for a system with unreliable resources is necessary. In our previous work, we proposed a robustness intensification algorithm based on a deadlock avoidance algorithm to ensure that processes not requiring unreliable resources can operate smoothly. In this article, a new robust method, which combines the advantages of absorbing policies and distributing policies with the aid of critical regions, is developed. It includes two algorithms, i.e., a robustness algorithm for processes not requiring unreliable resources and an intensification algorithm for processes requiring unreliable resources. Compared with existing absorbing and distributing policies, the new robust method can increase the production rate while decrease the algorithm complexity for a system allowing flexible routes apart from each process stage acquiring more than one type of resources.