RObot Supported Education for children with ASD

Children with autism often face challenges in developing effective communication and social skills. In the ROSA project, social robots are used as a tool to teach children with autism and developmental disabilities language and communication skills combined with social skills. Social robots have special characteristics, such as a human-like physical bodies, which distinguish them from other learning technologies. Many children with autism develop a strong interest for such robots, which can help maintain focus and attention, thereby providing better conditions for learning. The project is interdisciplinary and combines knowledge about the capabilities of social robots with expertise in special education focused on supporting children with autism, as well as competence in health and technology, digital learning design and ethics. We have used an iterative, cyclical and user centred approach involving pupils, staff and other stakeholders. Adapting the work to the everyday school context and real use situations has been important to promote relevance both after the project ends and for other schools in Norway. The ROSA project has placed strong emphasis on ethics throughout the research and development process. Experience from earlier projects has shown that formal ethical approvals alone do not always capture the ethical considerations that may arise when working with vulnerable groups. For this reason, we chose an approach where ethical questions were integrated into the iterative project work and discussed continuously in response to real situations, different disciplinary traditions and the varied needs of participants. A particularly important principle has been a process oriented understanding of consent, where children’s autonomy and opportunity to withdraw from activities were followed up throughout the development and evaluation phases, not only at the beginning. In this way, ethical reflection has been part of the ongoing work and has contributed to shaping everyday decisions in the project. To include the perspectives of children with limited communication abilities, the project has also explored the use of digital storytelling. The project has developed the ROSA solution, which consists of a repertoire of activities and tasks that can be carried out on social robots in combination with a tablet. The ROSA solution includes both fully developed games and tasks, as well as a remote control function that teachers may use flexibly. Teachers may combine elements from the repertoire to create learning sessions tailored to each child’s needs and interests. The aim is to contribute to variation, engagement and motivation for learning, and to provide schools with expanded learning resources that can support each child’s developmental goals. To evaluate the effects of the ROSA solution, pupils’ communication and social skills had to be assessed before and after using the ROSA repertoire. After comparing a range of assessment tools, the project chose the SCERTS model—Social Communication, Emotional Regulation, and Transactional Support. This model was translated and adapted to Norwegian professional terminology and to practices at Frydenhaug School. The ROSA solution was piloted in two rounds, with 5 children in spring 2024 and 15 children in autumn 2025. The first pilot focused on examining how the ROSA solution functioned in an authentic learning context. Several technical adjustments were made following this pilot. The second pilot, which lasted four weeks, also aimed to test and further prepare the evaluation approach for the main study, including the use of SCERTS and video analysis. The main study was carried out in winter 2025 over a two month period, comprising around 55 instructional sessions using the ROSA solution. A substantial body of material was collected, including digital stories, SCERTS assessments, teaching plans, video recordings of the sessions, focus group interviews with teachers and explorations of ethical questions. Part of this material has been analysed, and in the final phase of the project we are continuing the analysis, working with publications and planning a final seminar. The Norwegian Computing Center leads the project, drawing on its expertise in social robots and digital inclusion. The work is conducted in close collaboration with Frydenhaug School in Drammen Municipality, the University of South Eastern Norway, the University of Birmingham, and the companies Innocom AS and Cyberbook AS.

Children with autism spectrum disorder (ASD) are characterized by behaviors that pose challenges to developing communication and social skills. The project aims to use social robots to teach language skills combined with social and communication skills to children with ASD. The project will create and evaluate the RObot Supported education for children with ASD (ROSA) toolbox with three parts: Content Creator, for easily creating tailored one-on-one lessons for the unique needs of each child with ASD; Robot Software runs lesson content customized to the robot's capabilities; and Review, for following progress and input for the next lesson. Rosa toolbox targets no specific robot, but content will adjust to a robot's capabilities. The project employs a sociocultural approach with experts in social robotics, education, and ASD and conduct technological and interdisciplinary research in abstracting robot capabilities, robot-assisted language learning, customizing and personalizing technology for education, and education content for children with ASD. The final prototype will be evaluated in a year-long trial with over 50 children with ASD using the social robot in sessions at school several times a week. The children's language, social, and communication abilities will be compared to at least 50 children with ASD from previous years not in the trial. Items measured include children's motivation and engagement when working during the sessions. Quantitative and qualitative methods will be used and ethical challenges related to use of social robots for children with ASD will be addressed. Beyond helping children with ASD develop skills, the project will create knowledge about how the effect of embodiment, social robots, language learning, robot training, how social robots can improve language, communication and social skills of children with ASD. The toolbox can be further developed and commercialized to include other content and target other groups for learning.