Global Vs Local: A Comparison of Two Approaches to Perform Diagnosis in Networks of Mechatronic Agents

Natural systems often exhibit, at a macro level, a tendency to order that emerges from simple and local interactions between the system components. This idea that the whole can denote a consistent behavior that is more than the sum of individual components has inspired the design and development of many artificial systems. Recently, with the pervasiveness of IT in the automation domain the application of such principles became interesting as a mean to developed seamlessly plug gable, robust and fault-tolerant mechatronic systems. These concepts, however, have been mainly applied from a control perspective. This paper argues the regulatory significance of diagnosis in such complex systems and discusses two strategies (fully centralized with global information and fully decentralized with local information) to perform a diagnostic analysis that explores the network dimension of future mechatronic systems.