The research project is concerned with fundamental questions about the structure and composition of declarative knowledge bases for Companion systems and concentrates the symbolic modeling of information necessary for planning, decision making, interaction, dialog processing.

In this context, methods for systematically and consistently transferring situation and emotional information to the declarative knowledge representation level are being developed, focusing on correctly handling uncertainty.

A second key aspect lies on tool support for creating knowledge bases. These consist of several model components which must be coordinated and free of contradications, both each component alone and in their interplay. Special consistency criteria for the model components of a Companion system are defined for this purpose. Afterwards, appropriate mechanisms for checking these criteria are developed and implemented in a tool for systematic model construction.