Renée van den Heuvel (PhD) studied during her PhD research (2013-2018) the possibilities of robots to support play in children with severe physical or multiple disabilities. With two different robots (IROMEC and ZORA) she explored the potential of these robots for this target group in rehabilitation and special education. Therapists and teachers were closely involved in the development of interventions and the execution of robot sessions with individual children or in small groups. Results of the studies show that robots contribute in a promising meaningful way to the support of play and the achievement of goals in both rehabilitation and special education for children with severe physical or multiple disabilities. For example, robot ZORA showed promising results, especially in the role of instructor, rewarder or motivator. These new insights offer new chances for these children and the professionals involved in their education and rehabilitation.

Rianne Jansens (MSc. OT) investigated whether the PLEO robot could support well-being before, during and after intervention of children admitted to the hospital or undergoing an intense and long-term treatment. Pedagogical staff, employed at the paediatric ward to support well-being of children and parents in relation to the medical treatments, offered PLEO robot in a personalized way and for personalized well-being related goals such as a relief from stress, anxiety or boredom, a stimulus for physical (activity), enhancing interaction with other or play. Results showed that for six of nine children, the predetermined goal(s) were achieved. PLEO can be appealing to a child when offered in a minimal driven play situation because of its capability to accommodate different behaviour in children. This made the pedagogical staff perceive PLEO robot as a potential tool for interventions at the paediatric ward. Although PLEO has limited functionality, it might stimulate children in disadvantageous situations to play.

Brenda Littler (MSc) conducted a systematic review to explore and gather evidence on the use of social robots in hospital settings intending to reduce anxiety or distress in children under the age of 18. There were eleven papers from the year January 2009 until July 2020 that were discovered, across eight databases. Overall, the evidence implied that social robots do have the potential to reduce negative emotions, such as anxiety and distress in children visiting the hospital. The findings from the systematic review have led to another study which is planned to start in the year 2021. It will investigate the types of interactions that occur between a child and a robot while in hospital. The idea is to introduce three social robots (Pepper, Nao and Miro) into five different hospital settings at the Sheffield Children Hospital and explore the types of interaction and play that occur, as well as the emotional impact. The research team will gather information on which social robot interacts effectively with children in the hospital and which settings have the feasibility to have a social robot in their department. The final goal of the project is to evaluate the potential of introducing one of these three robots into one of those five settings, to reduce anxiety and distress.

Claire Huijnen (PhD) has studied the potential of robot-mediated interventions in and with educational practices for children with ASD. Professionals in the field expected that robot KASPAR would be able to contribute to many objectives that professionals work on for children with this heterogeneous disorder, such as imitation in play, making contact, turn taking, attention or developing interest in play. To support professionals to integrate a robot-mediated intervention with KASPAR tailored to the child’s objectives, a template was offered for the development and implementation of the intervention in the daily practice. A pilot study analyzed micro behaviours of children with ASD at a special school interacting with KASPAR. Results indicated that children reacted positively and freely to the robot, they made contact both verbally and non-verbally, showed initiative and had a long attention span. The use of robots for children with ASD seems a promising direction to augment or complement (special) education.

Alessandro Di Nuovo (ADN) is Professor of Machine Intelligence at Sheffield Hallam University (SHU) and leader of the research theme on technological and digital innovations to promote independent lives at the Advanced Wellbeing Research Centre. ADN has an extensive track record of leading interdisciplinary research and innovation. ADN has been leading projects in fundamental and applied topics in AI and Robotics, funded by public (EU H2020, EPSRC), private (IBM, NVIDIA) and charities (Sheffield Children’s Hospital). Currently, ADN is a lead scientist (SHU PI) of the H2020 MSCA-ITN European Training Network on PErsonalized Robotics as SErvice Oriented applications (PERSEO, 2021-24), a multidisciplinary project to train the new generation of interdisciplinary researchers and professionals for the forthcoming market of personal robots. ADN was the PI of the EPSRC project NUMBERS (2017-19), which delivered the first step toward his vision by creating a developmental neurorobotics model of number understanding. The research evidenced similarities between children and robotic embodiment in the early development of numerical cognition (Di Nuovo & McClelland, Nature Machine Intelligence, 2019). ADN won multiple a IBM Awards (total cash $60,000) for a prototype of a robot-led psychological cognitive assessment system powered by the IBM Watson Cloud AI services. ADN coordinated the H2020 MSCA-IF CARER-AID (2017-19), which developed an intelligent robotic assistant to support caregivers in early diagnosis and treatment of individuals with autism spectrum disorder associated with intellectual disability, in collaboration with a specialised healthcare provider. Previously, ADN was leading the human-robot interaction in the FP7 Large Scale project ROBOT-ERA (2012-15), a multidisciplinary international project concerning the implementation and evaluation of several advanced robotic services for supporting the elderly.

Andrea Bonarini (PhD) investigated the possibility of adopting both general-purpose autonomous robots, and robots designed for special purposes in play-like activities with people affected by neuro-developmental disorders (NDD). With his team, he developed different versions of an autonomous robot (Teo) that can react to actions done by the interacting people. The therapist can also control it remotely, so exploiting a sort of avatar that can play a role the therapist cannot usually play (e.g. crying when attacked). Now that Deep Learning capabilities can be embedded in low-cost devices, we are exploring activity recognition from a camera, and sound recognition and analysis from a microphone array, to improve the type of interaction the robot can do autonomously. Following suggestions by therapists and subjects, in a co-design activity, Andrea Bonarini’s team also developed more than thirty autonomous robots for specific games. Among the others, we mention a monkey plush, sensible to movements and touch, and able to participate to a round-the-rose circle: when the circle closes it starts to sing the typical song. Many other robots have been developed (a partial collection is available from http://playbot4all.polimi.it) in the spirit of involving people to develop, or at least exploit, skills, and, at the same time, play, since play is a fundamental activity for developing abilities, often missing for this people either because of inadequacy of the available toys, or because their time is dedicated to structured therapies. Bonarini’s team is also exploring the possibility of objective evaluation of the performance and progresses from sensors on board of the robot, in particular manipulation. They developed systems to interpret signals from accelerometer both to monitor how a toy is used and to prevent possible misuses of it (e.g., throwing).

Lorenzo Desideri (PhD, PsyD) has investigated the use of a humanoid robot (NAO) to increase the effectiveness of treatment interventions targeting children with ASD. The findings of the studies conducted suggest that social robots may be a useful tool to increase the engagement of children with ASD in treatment interventions conducted in well-controlled clinical settings. To face with the challenge of implementing robot-based activities in less restricted environments such as inclusive school settings, he and his team have recently developed RoBò, a service delivery approach aimed at supporting teachers in the design, implementation and conduction of educational group-based activities performed using the NAO robot. The aim of such group-based activities is to increase the opportunities for social interaction of students with ASD to improve their social responsiveness (e.g. verbal initiations, eye contact).

Pedro Encarnação (PhD) has been involved in research projects exploring the potential of robots as tools (assistive technologies) to support manipulation for children with neuromotor impairments. By controlling a robot through an adequate interface, children with physical impairments may be able to manipulate objects thus gaining access to play and academic activities that require manipulation. In those projects, physical Lego® Mindstorm® robots have been adapted to support children with neuromotor impairments integrated in regular schools in performing mathematics, language, and science & social studies academic activities along with their peers. Virtual systems in which children controlled a virtual robot to manipulate virtual objects, performing the activities on a computer screen, were also developed. Since often children with neuromotor impairments use speech generating devices for communication, robots can be controlled through those devices building on the access skills already acquired by the child. Both the physical and virtual systems have shown to be useful resources to support participation of children with neuromotor impairments in classroom activities. A common outcome in studies where children with disabilities use robots as assistive tools, is that adults are surprised by the level of accomplishment of children with disabilities, as they often underestimate the abilities of children with disabilities. In fact, it has been shown that the use of augmentative manipulation assistive robots to perform specially designed tasks may provide a proxy tool to assess young children with disabilities’ cognitive skills

Anna Lekova, (PhD) investigated whether the NAO robot could aid the auditory-verbal therapy sessions of children with communicational disorders, either typical - with hearing problems or children with ASD and speech impairments. The unique characteristic of the technology, NAO does not have a human mouth and therefore does not allow lip-reading or other non-verbal cues, had been successfully used from Ioannou et al. (2019). They found that NAO create playful and engaging auditory-verbal therapy sessions, however they lack in quantitative tools for assessing the progress in the development of the auditory and verbal skills. In the ongoing H2020 Project CybSPEED we explore how programmable robots may also be useful as technologies for quantitative assessment of the progress in the rehabilitation. Six kind of rehabilitation games were developed and deployed on NAO, one of which is the Ling Six-Sound Test. Thus, the therapist is focusing only on the individual learning skills, while the assessment of the children attentiveness and auditory-verbal skills is obtained in a quantitative and technology based manner online or post-hoc. The exploited by Lekova et al. (2019) idea behind the Internet of Things (IoT) for integrating robots and intelligent sensors in device and platform independent way, allowed the therapy session to be performed and assessed remotely.

Luc de Witte (PhD) supervised many studies with different robotic platforms with users of different ages and health conditions in cooperation with informal and formal caregivers in the real-life contexts of users.