Cognition-based self-optimisation of an automotive rear-axle-drive production process

The production of automotive rear-axle drives is a complex process. This is due to many involved process steps, factors and interdependencies between processes, materials, means of production and individuals acting in this environment. In general their effect on product variations is not fully comprehended. Hence, a holistic analytical model is only possible in parts of the production. In this paper a modular approach is presented to make the production more flexible and enable it to react faster on product variations. This is achieved by a Cognitive Production System (CPS), which is based on accumulating, storing and processing of process knowledge so that it can be applied to similar cases. Through the combination and interaction of Cognitive Tolerance Matching (CTM) and Agent-based Systems the performance of the CPS is enhanced. The work discusses the set-up of such a CPS for the production of automotive rear-axle-drives with the focus on the failure state agent.