From Interaction Science to Cognitive Interaction Technology

A cascade of revolutions has transformed robotics into a new science that begins to link physical concepts of control and interaction with qualities and concepts analyzed so far mainly in disciplines such as psychology, biology, linguistics or the social sciences. A visible manifestation of this transformation is the rapid evolution of robot body morphologies within the last decade: robots with different degrees of humanoid appearance -ranging from a coarse analogy of humanoid body structure to amazingly detailed replications of human appearance-are now being developed in numerous labs and not few of them have become commercially available. This has completely reshaped the interface between robots and people: expressive faces require a rethinking of the way robots and humans can interact. Human-like hands offer new levels of interaction provided we can find ways to coordinate their rich degrees of freedom to bring their skills closer to what we see in humans. And progress in processing power, storage density and sensing technologies allows to create and maintain rich representations to cope with contact at all levels from the physical to the social, thereby turning robots from the “contact minimizers” of the past into interactive agents that can assist people in natural ways and in natural environments. These developments have shaped much of CITEC's vision of cognitive interaction technology and the associated research agenda along the four topic pillars of motion intelligence, attentive systems, situated communication, and memory and learning. We describe some of the underlying methodology and how we have mapped it into a dedicated lab infrastructure to investigate cognitive interaction processes across a wide spectrum of spatial and temporal scales. We present selected examples of cognitive interaction that are in the focus of major CITEC research lines: the Bielefeld expressive robot head Flobi that offers a versatile platform to find new ways to make human robot interaction more natural and to enrich it with new layers of usefulness; the advancement of touch as the modality where social and physical contact come together, for instance, through sensors that are more skin-like and ways of how to use such sensors for haptic exploration of surfaces or as research tools for studying human haptics; and interdisciplinary work towards a deeper understanding of “manual intelligence” and its replication in bimanual robots to enable them to flexibly manipulate everyday objects and to cooperate more naturally with people. Finally, we report on the status of the newly started project FAMULA that connects 10 CITEC groups to study and replicate the process of familiarization with novel objects through combined use of embodied manipulation and language. Turning to the perspectives of our field into the future, we argue that we can take inspiration from the systematic elucidation of principles and mechanisms of interaction that enabled physics to advance to a new level. Taking guidance from this success story may put us on a similarly grand path: creating a science for elucidating the inner workings of cognitive interaction that arises when we take the step from physical particles to cognitive agents and using our insights to create better bridges between humans and tomorrow's technology.