Projekt CommRob

This project addresses the introduction of robots in everyday human environments and their close cooperation with people. The major focus in this regard is to enable a robot to execute tasks while communicating with a user, also in the context of other robots operating in the same human environment. Such everyday human environments may be dynamic and complex like shopping centres and airports. Many current projects (Cogniron, PapeRo, etc.) design and develop personal or companion robots, operating in close interaction with a user at home, i.e., in a small and static environment. A major challenge of CommRob is to demonstrate that companion robots can perform services efficiently and safely in a complex and cluttered dynamic environment, shared by other humans and other robots. To share such an environment with humans and to cooperate with them requires certain communication abilities from the robot, but also from the humans. For an introduction of robots in everyday human environments it will not be possible to train all the humans involved. Based on this assumption, we postulate that the robot will have to possess certain communication abilities like humans. It must be intuitive for humans without special training to communicate with such a robot, e.g., by approaching it, by touching it or via simplified speech. To share such an environment with other robots as well, even complicates matters further for a given robot. We assume that all the robots in such an environment will have to possess the same communication abilities. This may facilitate robot-robot communication on the same basis as human-robot communication. Therefore, we will pursue a unified communication approach. In addition, everyday human environments are dynamic and will are challenging in robotics. To share such environments, advanced behaviours of robots are required. We will especially address issues of moving safely in a dynamic environment, i.e. how to avoid dynamic obstacles, how to perform efficient self-localisation, to support users or cooperate with other robots. We will study these issues with a personal robot in mind that will primarily perform multimodal perception and interpretation to communicate with its current user. This robot will also have to communicate with other robots and interact with other humans sharing the same physical space.

The high-level objectives of this project are twofold: they concern a unified approach to robot- robot and human-robot communication, and the validation of these concepts through the integration of advanced robotics behaviours.

A primary objective is then to achieve high-level multimodal communication with and among robots. In particular, certain robot behaviours will be acts of communication. For example, the movement of a robot may mean a path traversal or an attempt to communicate, or both. This is analogous to the movement of a human in a human environment, and should be dealt with analogously for robots.

The other primary objective concerns advanced behaviour of robots in everyday human environments. This involves the technical specification of single behaviours, their fusion into a network and their integration with the interaction and communication subsystems. The resulting control architecture should be modular and open. Moreover, since humans share the same physical space with such robots, safety is an absolute requirement for their behaviour, especially their movements. Several behaviour classes will be implemented and will lead to scalable performance of the robots: from basic safety, error recovery, local behaviours (e.g., path planning, dynamic obstacle avoidance, self-localisation) up to advanced (e.g., user supporting and following) and robot-robot-cooperative behaviours (e.g., traffic rules, negotiation, leader following, cooperative map building). Both the robot or the user will act accordingly to switch between modalities depending on the environmental context or on the robot status.

Our partners in the consortium of CommRob are the TUW (Vienna), KTH (Stockholm), LAAS (Toulouse), etp (Freiburg) and Zenon (Athens).

• Gefördert: EU
• URL: http://www.commrob.eu
• E-Mail: This e-mail address is being protected from spambots. You need JavaScript enabled to view it
• beteiligte Forschungsbereiche: IDS