Research in robotics has traditionally emphasized low-level sensing and control tasks including sensory processing, path planning, and manipulator design and control. Research in Cognitive Robotics, on the other hand, emphasizes those cognitive functions that enable robots and software agents to reason, act and perceive in changing, incompletely known, and unpredictable environments. Such robots must, for example, be able to reason about goals, to choose actions and to focus on patterns, objects and events according to the task execution and the cognitive states of other agents, by taking into account time, resources, and the consequences of their decisions. In short, cognitive robotics is concerned with integrating reasoning, perception, and action within a uniform theoretical and implementation framework.

The term cognitive robotics and the vision that knowledge representation and reasoning plays a fundamental role in the design of cognitive robots was first laid out by the late Ray Reiter in his lecture on receiving the Research Excellence Award by the International Joint Conference on Artificial Intelligence (IJCAI) in 1993. Since 1998, biannual Cognitive Robotics workshop with Dagstuhl being the seventh in this series.

While the earlier workshops were largely a forum for presenting state-of-the-art research results, the purpose of the Dagstuhl event was to broaden the view and bring together people from various disciplines to shed new light on the issues in cognitive robotics. In this respect we were very fortunate to have participants from areas such as robotics, machine learning, cognitive vision, computational neuroscience, and knowledge representation and reasoning.

Given the diversity of the group, we spent the first day with tutorial-style presentation, starting out with an overview of Cognitive Robotics in the sense of Ray Reiter's vision by Gerhard Lakemeyer. This was followed by presentations on Computational Neuroscience by Laurent Itti, Planning and Execution Monitoring by Brian Williams, Probabilistic Reasoning by Eyal Amir, Cognitive Vision by Jim Little, and Human-Robot Interaction by Geert-Jan Kruijff. The rest of the Workshop consisted of research presentations, a panel, and three breakout discussion groups on the following topics: the nature of perception, symbolic and numerical uncertainty, and the role of automated reasoning.

The Proceedings collected here represent submissions from all but 2 of the participants at the Workshop in the order they were given. In some cases we have short abstracts of the talks, and in others, full research papers. We thank the participants and the copyright holders for permission to use these papers.