Robust control of automated manufacturing systems with complex structures using Petri Nets

Deadlock problem is a cutting-edge research subject in automated manufacturing systems (AMSs). Most work to date assumes that the system resources are reliable. However, an AMS is composed of plenty of numerically-controlled machines interacting with each other. Resource failures occur unexpectedly especially due to increasingly complex system structures, e.g., AMSs with assembly operations and flexible routes. Thus, resource failures will be taken into consideration in this paper. Through dynamic analysis and real-time control, our robust supervisor can automatically allocate resources so that processes blocked by unexpected resource failures do not hinder the production of parts not necessarily requiring failed resources at any time. Our approach achieved in a distributed and dynamic way is different from conventional methods based on structure-oriented and monolithic specifications.