Design framework for the integration of cognitive functions into intelligent technical systems

The integration of cognitive functions will enable mechatronic systems to be superiorly embedded into their environment and to follow their system objectives independently. The intention is to develop self-optimizing systems, which can optimize their behavior by themselves to become more flexible, robust and user-friendly. Numerous challenges, however, become apparent on the way to such intelligent technical systems. The development is characterized by an increasing involvement of non-technical disciplines like cognitive science, higher mathematics or neurobiology. Existing design methodologies are focusing technical disciplines on the one hand and non-technical disciplines on the other hand. For instance, there is a lack of a systematic coupling of those disciplines, which are relevant for the exploration of cognitive functions, with the general engineering approach in product development. To rise to these challenges, the integration of cognitive functions has already to be supported with some kind of methodology. Focus of the methodology must be the early stages of the development. Within this design phases the developer have to modify the principle solution in common. Hence, important requirements occur in terms of the intensified interdisciplinarity of the development and the increasing system complexity. Therefore, a design framework for the integration of cognitive functions into self-optimizing systems has been developed which integrates both, existing and newly developed methods in a well-structured procedure. For this purpose, in section two, we will introduce the concept of self-optimizing systems and the operator-controller-module. Afterwards we will describe the need of action in section three and the state of the art: “design framework for cognition” in section four. In section five, we present our developed design framework for the integration of cognitive functions into intelligent technical systems. Therefore, we will explain the procedure model and a specification technique to describe self-optimizing systems. In addition, we will present a uniform type of solution patterns for the reuse of once successfully implemented knowledge and the solution pattern knowledge base for the tool support. To conclude, we will sum up the major points and give a short outlook on our future work.