Robots that understand and interact
Traditional robots that are used in industry often lack any perceptual abilities as they are programmed to perform only one thing. However, to be useful for other real-world applications, people should be able to easily train and control them. With EU funding, the 'Interactive robotics research network' (INTRO) project laid the necessary foundations for robots to break into new markets by creating novel systems with cognitive and multi-modal interaction functions. Through the Marie Curie Initial Training Network (ITN) programme, the project contributed to creating a new generation of scientists focusing on the technologies needed to build intelligent robots. Such robots were expected to function in close interaction with humans in unstructured, real-world changing conditions. Project members recruited eight early-stage and two experienced researchers who explored such themes as cognitive psychology, sequence learning and imitation as well as behaviour and intention recognition. Other research topics included dynamic gesture analysis, emotional interaction, robot safety and interaction, failure detection and recovery, and intelligent interface design. Except for the individual research projects, the programme also included a series of seminars, network-wide workshops and courses to offer researchers a strong theoretical background. Project activities resulted in two demonstrators: a robot waiter, and an urban search-and-rescue robot. INTRO findings were disseminated through the project website, a video presentation and publications in several journals.
Keywords
Robot, human–robot interaction, interactive robotics, cognitive, multi-modal interaction, intelligent robot, robot waiter, search-and-rescue
