Supervising Industrial Distributed Processes Through Soft Models, Deformation Metrics and Temporal Logic Rules

Abstract

Typical control solutions for future industrial systems follows a top-down approach, where processes are completely defined at high-level by prosumers (managers) and, later, the control infrastructure decomposes, transforms and delegates the execution of the different parts and activities making up the process into the existing industrial physical components. This view, although may be adequate for certain scenarios; presents several problems when processes are executed by people (workers) or autonomous devices whose programming already describes and controls the activities they perform. On the one hand, people execute processes in a very variable manner. All these execution ways are valid although they can be very different from the original process definition. On the other hand, industrial autonomous devices cannot be requested to execute activities as desired, and their operations can only be supervised. In this context, new control solutions are needed. Therefore, in this paper it is proposed a new process supervision and control system, focused on industrial processes executed in a distributed manner by people and autonomous devices. The proposed solution includes a soft model for industrial processes, which are latter validated through deformation metrics (instead of traditional rigid indicators). Besides, in order to guarantee the coherence of all executions, temporal logic rules are also integrated to evaluate the development of the different activities. Finally, an experimental validation is also provided to analyze the performance of the proposed solution.