PUB - Publikationen an der Universität Bielefeld

A characteristic feature of intelligent systems is the interaction of numerous processes in areas as perception, attention control, sensori-motor interaction, memory, reasoning, or speech understanding. While in the past the primary goal in building artificial cognitive systems was to realize a suitable number of respective partial functions, the focus in this area has recently shifted more and more towards investigating overall architectures for integrating a large number of heterogeneous partial skills to make them interact intelligently. The development of such architectural concepts is one of the main goals of the Bielefeld Collaborative Research Center (CRC) 360 Situated Artificial Communicators. In this contribution, results for three partial demonstrators are presented, which focus on three central aspects of system integration: perception, cognition, and learning. These demonstrators have also been combined and work on a realistic level of complexity. Consequently, we believe that the concepts developed can contribute essentially to the research on architectures of artificial cognitive systems.