Are Robots Ready to Deliver Autism Interventions? A Comprehensive Review

This article presents a review of the contemporary robotics research with respect to making robots and human-robot interaction (HRI) useful for autism intervention in clinical settings. Robotics research over the past decade has demonstrated that many children with autism spectrum disorders (ASD) have a strong interest in robots and robot toys and can connect with a robot significantly better than with a human. Despite showing great promise, research in this direction has made minimal progress in advancing robots as clinically useful for ASD intervention. Moreover, the clinicians are generally not convinced about the potential of robots. A major reason behind this is that a vast majority of HRI studies on robot-mediated intervention do not follow any standard research design and, consequently, the data produced by these studies is minimally appealing to the clinical community. In clinical research on ASD intervention, a systematic evaluation of the evidence found from a study is performed to determine the effectiveness of an experimental intervention (e.g., a robot-mediated intervention, RMI). An intervention that produces a stable positive effect is considered as an evidence-based practice (EBP) in autism. EBPs enable clinicians to choose the best available treatments for an individual with ASD. The ultimate goal of RMI, therefore, is to be considered as an EBP so that they can actually be used for treating autism. There are several criteria to measure the strength of evidence, and they are mostly geared toward rigorous research design. The research on RMI, therefore, needs to follow standard research design to be acceptable by the clinical community. This paper reviews the conM. Begum and H. A. Yanco Computer Science, University of Massachusetts Lowell E-mail: {mbegum,holly}@cs.uml.edu R. W. Serna Psychology, University of Massachusetts Lowell E-mail: Richard [email protected] temporary literature on robotics and autism to understand the status of RMI with respect to being an EBP in autism treatment. First, a set of guidelines is reported which is considered as a benchmark for research design in clinical research on ASD intervention and can easily be adopted in HRI studies on RMI. The existing literature on RMI is then reviewed with respect to these guidelines. We hope that the guidelines reported in this paper will help the robotics community to design user studies on RMI that meet clinical standards and thereby produce results that can lead RMI toward being considered as an EBP in autism. Note that the paper is exclusively focused on the role of robots in ASD intervention/therapy. Reviews on the use of robots in ASD diagnosis are beyond the scope of this paper.