Towards Multi-Modal Interaction with a Mobile Robot

Off-the-shelf mobile robot platforms with good navigation capabilities can nowadays be used for the exploration of advanced human-robot interfaces. The development of systems with natural interaction capabilities is an important prerequisite for the widespread use of robots in home and office environments. In this paper, we present the Bielefeld Robot Companion (BIRON), which uses a multimodal approach for human-robot interaction. Visual cues, laser range data, and acoustic data are used in conjunction with a data association framework for detecting and observing humans. For interaction with people, an attention mechanism is defined that enables the robot to detect communication partners and provide appropriate feedback.