Guest Editorial

1. Getting to Know Socially Intelligent Robots

Human-Robot Interaction and Communication is a quickly growing research area at the intersection of research fields such as robotics, engineering, psychology, ethology and cognitive science. Significant initiatives are currently underway funded by public, academic, governmental as well as industrial initiatives, exploring and aiming at advancing this research field and opening up novel and challenging applications. Robots moving out of laboratory and manufacturing environments face hard problems of perception, action and cognition. For robots to be accepted as assistants or companions in people's private homes and everyday environments technological solutions do not suffice: The ‘human in the loop’, as the potential customer and user will decide on the ultimate success of a ‘home robot’ as a product. Application areas that heavily involve human contact are a particularly challenging domain.

Human societies have easily assimilated new technologies, such as mobile phones, but it is less clear in which application areas robots will be accepted. Robots as embodied beings, physical, possibly humanoid or android entities that share our living environments and accompany us in our lives will have a certain degree of autonomy, initiative, cognitive skills and will communicate and interact with people in ways inspired by human-human contact. Interaction and communication of embodied physical robots with humans is multi-modal, and involves deep issues of social intelligence, communication and interaction that have traditionally been studied primarily in psychology and other areas. The design of a robot's behaviour, appearance, and cognitive and social skills is highly challenging, and requires interdisciplinary collaborations across the traditional boundaries of established disciplines.

2. The IEEE Robot and Human Interactive Communication Symposium Series

IEEE Ro-Man 2006 provided a forum for an interdisciplinary exchange for researchers dedicated to advancing knowledge in the field of human-robot interaction and communication. Importantly, Ro-Man has traditionally adopted a broad perspective encompassing research issues of human-machine interaction and communication in networked media as well as virtual and augmented tele-presence environments. Submissions of the latest cutting-edge work were invited from a variety of research areas that can advance our understanding of human-robot interaction and communication, including areas of engineering and information sciences as well as psychology, social sciences, cognitive science and related areas.

The annual Ro-Man International Workshop series originated in 1992, with the first workshop held at Hosei University in Japan. Since then, different Japanese, European and USA institutions have hosted the workshop. In 2006 IEEE Ro-Man become a Symposium and was hosted in United Kingdom by the Adaptive Systems Research Group at University of Hertfordshire. The event included tutorials, special organized sessions, distinguished invited keynote speakers, a robot design contest, and awards for best papers. Relevant topics for this vibrant area of Human-Robot Interactive Communication include but are not limited to:

innovative robot designs for human-robot interaction (HRI) research

3. Selected Papers

Three papers are included in this special section. They have been selected on the basis of recommendations in referee reports from the conference indicating suitability for expansion to journal papers and based on relevance to this journal. The authors were invited to substantially extend and revise the papers in an iterative process of anonymous peer review by expert referees. (Other papers from IEEE Ro-Man 2006 were selected for a special issue of the journal Interaction Studies: Social Behaviour and Communication in Biological and Artificial Systems).

The lead article «Self-imitation and Environmental Scaffolding for Robot Teaching» by Joe Saunders, Chrystopher L. Nehaniv, Kerstin Dautenhahn, and Aris Alissandrakis reports on new scalable methods for scaffolding human-robot social learning and teaching. In humans, evolutionary predecessors to the sophisticated capacity to learn from observational imitation may have involved self-imitation where an agent avoids the complexities of mapping what it sees another do to its own body (the famous «correspondence problem») by learning and replicating actions it has experienced through the manipulation of its body. The work introduces and evaluates a novel robot social learning system using self-imitation that is inspired by psychological models of motor control and ideas from social scaffolding seen in animals. With scaffolding by one or more human teachers, robot competencies can be built by constructing hierarchical state-action memory maps of the robot's interaction within that environment. This scaffolding process provides a mechanism to enable learning to be scaled up. The resulting system allows a human trainer to teach a robot new skills and modify skills that the robot may already possess. Also, the system allows the robot to notify the trainer when it is being taught skills it already has in its repertoire and, moreover, to direct and focus its attention and sensor resources to relevant parts of the skill being executed. This architecture is validated on a physical robot that is taught scaffolded sets of new skills. The contribution «Situated Dialogue and Spatial Organization: What, Where… and Why?» by Geert-Jan M. Kruijff, Hendrik Zender, Patric Jensfelt, and Henrik I. Christensen reports on conceptual spatial mapping using situated dialogue via studies on an human-robot interaction architecture for «human-augmented mapping», implementing various interaction strategies observed in Wizard-of-Oz studies, and discusses an ontology-based approach to multilayered conceptual spatial mapping. This helps provide a common ground for human-robot dialogue in order to bridge the divide between the robot's sensor-based representation of the world and meaning extracted from human speech utterances, and leads to establishing spatial references in situated dialogue between human and robot about their environments. The work contributes towards the goal of situational awareness by robots interacting with humans in complex environments.

Finally, the article «A Monocular Pointing Pose Estimator for Gestural Instruction of a Mobile Robot» by Jan Richarz. Andrea Scheidig, Christian Martin, Steffen Müller and Horst-Michael Gross presents work relevant for gesture recognition in highly interactive mobile companion robots using multi-modal person detection and tracking. The authors present a hierarchical neural network architecture that estimates a target point indicated by a pointing human interaction partner. They show that even using poor quality monocular image data and low-cost cameras that an accuracy better than that of a human viewer of the same data can be achieved in a user-indepedent manner.

The guest editors are grateful to the anonymous reviewers of these papers for their conscientious and valuable work, and to Tony Belpaeme for invaluable assistance with the refereeing process.