. Increasingly socially intelligent agents (software or robotic) are used in education, rehabilitation and therapy. This paper discusses the role of interactive, mobile robots as social mediators in the particular domain of autism therapy. This research is part of the project AURORA that studies how mobile robots can be used to teach children with autism basic interaction skills that are important in social interactions among humans. Results from a particular series of trials involving pairs of two children and a mobile robot are described. The results show that the scenario with pairs of children and a robot creates a very interesting social context which gives rise to a variety of different social and non-social interaction patterns, demonstrating the specific problems but also abilities of children with autism in social interactions. Future work will include a closer analysis of interactional structure in human-human and robot-human interaction. We outline a particular framework that we are investigating. 1

This paper proposes a technique for quantitatively describing and analysing robot-human interactions in terms of low-level behavioural criteria (so-called micro-behaviours). In order to demonstrate the usefulness of this technique, we describe a case study that was conducted as part of the Aurora project where we develop robotic toys as therapeutic tools for children with autism. In this project we made explicit choices of how to assess robot-human interaction and how to study the impact of interaction. Results of a comparative study with autistic children are shown where we focus on eye gaze behaviour. The results point out common tendencies as well as clear differences among the children, important information for future development of robots in the Aurora project. We propose that this technique is applicable to a wide range of application areas that involve robot-human interactions. The work presented in this paper is intended to open up a discussion on appropriate techniques to systematically assess robot-human interactions. Such research is important for the development of robots in human-inhabited environments. 1

Software developers, designers and researchers have been looking to technology for solutions to help and educate people with autism for over two decades. There are many examples of seemingly successful technology-based products and prototypes, yet very little is known about how well these solutions are currently integrated into lives of children and adults with autism and their families. This paper reports on results from an anonymous on-line survey intended as a first step to elucidate information about software and technology use. Additionally, data was analyzed to aid creation of future technology-based products for people with autism that are not just effective, but that also meet important user goals and align to their interests and strengths. Major findings included: (1) very few respondents (25%) had any experience with software or technology designed for people with cognitive disabilities; (2) when asked an open-ended question about what they desire in technology design, respondents reported three major goals (social skills, academic skills, and organization skills), and many suggestions for improvements to software and hardware design; and (3) technology was reported as both a major strength and interest for people with autism.

This paper reports results deriving from the Aurora project (www.aurora-project.com) where we have pioneered research into the possible use of robots in autism therapy. Autistic children have di#culties in social interaction, communication and fantasy and imagination. As part of the project we run trials where autistic children are playing with a small, non-humanoid mobile robot that can engage children in simple interaction games. In our project we focus on the behavioural, rather than the a#ective level of robots used in therapy, cf. [12]. In this paper we first discuss in more detail varieties of interactions where one child, or two children simultaneously, play with a robot. We then outline a new research direction in the project which studies how a mobile robot can adapt to individual children. Quantitative examples of activity levels in child-robot interactions are included. The paper concludes by outlining future research directions for adaptive robots in autism therapy.