Structured Information Processing for Self-Optimizing Mechatronic Systems

Self-optimizing mechatronic systems are based on intrinsic controller systems whose complexity by far exceeds that of currently available systems. In addition to procedures taken from artificial intelligence, procedures for a reconfiguration by means of appropriate design methods have to be integrated to fully implement self-optimization features. Special importance falls to a networking of such complex controller systems for the support of collaborative and emergent self-optimization. One main challenge lies in the safety-critical nature of the systems that requires the resulting software along with the technical system to show a predictably correct behavior in spite of networking, reconfiguration, and integration of procedures from artificial intelligence. The paper presents a concept for structuring and designing reconfigurable controller systems.