Workshop Description

The Cognitive Robotics Workshop is a prestigious international workshop that takes place every two years. It is normally co-located with one of the main AI conferences or events. Although not a rule, we like to switch the venue between North America and Europe. The previous workshops have been:

Research in robotics has traditionally emphasized low-level sensing and control tasks including sensory processing, path planning, and manipulator design and control. In contrast, research in cognitive robotics is concerned with endowing robots and software agents with higher level cognitive functions that enable them to reason, act and perceive in changing, incompletely known, and unpredictable environments. Such robots must, for example, be able to reason about goals, actions, when to perceive and what to look for, the cognitive states of other agents, time, collaborative task execution, etc. In short, cognitive robotics is concerned with integrating reasoning, perception and action with a uniform theoretical and implementation framework.

The use of both software robots (softbots) and robotic artifacts in everyday life is on the upswing and we are seeing increasingly more examples of their use in society with commercial products around the corner and some already on the market. As interaction with humans increases, so does the demand for sophisticated robotic capabilities associated with deliberation and high-level cognitive functions. Combining results from the traditional robotics discipline with those from AI and cognitive science has and will continue to be central to research in cognitive robotics.

This workshop aims at bringing together researchers involved in all aspects of the theory and implementation of cognitive robots, and discussing current work and future directions. While all aspects of cognitive robotics are of interest to the workshop, we especially welcome discussions and demonstrations of implemented systems.

Why the workshop is of current interest

We are facing a rapidly increasing number of computer systems that are embedded in physical devices, equipped with sensors such as web cams and RFID tag readers, have access to smart sensor networks, are combined through communication networks such as wireless LAN but also through semantic grids of computation modules, and they are connected to the World Wide Web, today's most comprehensive information source. The degrees of freedom of a single state-of-the-art dexterous hand exceeds those of many autonomous mobile robots. The capability of sensing devices, the computational infrastructure, and the advent of new actuator mechanisms suggest the realizability of unprecedented intelligent and physically realized application systems.

However, main bottlenecks and obstacles in realizing them are in perception, knowledge extraction from sensor data streams, action selection, system complexity. This increase of system complexity also causes the cost of system development and maintenance to explode, the required lifespan of the systems to be longer, the systems to become more fragile, and the interaction with the systems to get harder.

These problems cannot be coped with by merely improving the software engineering techniques. Cognition in technical systems, such as autonomous service robots, intelligent sensor-equipped environments, and peer-to-peer software office assistants, is currently an important focus of research in intelligent systems.

The current interest and the urge for research results is, among other things, demonstrated by a number of national and international research initiatives including the Darpa cognitive systems initiative of the Information Processing Technoloy Office, various integrated projects funded by the European Community, the British Foresight program for cognitive systems, huge japanese research efforts, for example in the context of Riken, to name only a few.

Another reason for the current standing of the field is that in recent years scientific disciplines studying cognitive systems have crossfertilized each other in various ways. Cognitive psychologists have found convincing empirical evidence for the use of Bayes estimation and cost function enabled control mechanisms in natural motion control. Bayesian networks and the associated reasoning and learning mechanisms have inspired research in cognitive psychology, in particular the formation of causal theory in young children. Also functional MRI images of rat brains localizing in mazes have shown neural activation patterns of place cells similar to multimodal probability distributions in robot localization using Bayesian filters.

These examples suggest that (1)~successful computational mechanisms in artificial cognitive systems tend to have counterparts with similar functionality in natural cognitive systems; and (2)~new consolidated findings about the structure and functional organization of perception and motion control in natural cognitive systems indicate much better ways of organizing and specifying computational tasks in artificial cognitive systems.