Using intelligent robots to detect body language and improve social development in children with autism spectrum disorder

Socially assistive robotics (SAR) is an expanding field within special education, proven to benefit the social development of children with autism spectrum disorder (ASD). This population often faces challenges in social interaction, particularly in understanding body language, leading to feelings of isolation and impacting their social lives. The use of robot-agent for practice is increasingly being adopted to address these challenges. In this study, an intelligent robot system was designed to enhance social interactions, particularly focusing on body language, among children with ASD. The system integrated an intelligent robot with a Kinect sensor to detect and execute body language actions, utilizing these developed modules for detecting body movements. This study explored the effects of the developed NAO-Recognitions of Body Languages and Imitation (NAO-RBI) system and analyzed the performance of three participants with ASD in these targeted body languages. The body-language gestures involved: raising a hand, high-five, handshake, "Please give me," "Come on!," arms akimbo with an angry face, and "Please!". A multiple-probe design was used to monitor improvements in the targeted actions, with a scoring module developed to gather scores from each participant. Results showed that the NAO-RBI system successfully recognized and demonstrated the seven body language gestures. This highlights the potential of SAR, combined with Kinect technology, to enhance body language skills in children with ASD. Comparing skeletal point data captured by Kinect with predefined standard body skeleton coordinates effectively executed the module functions. Robot-assisted practice proved to be a valuable tool for improving body language understanding and use for social interactions in this population. Further research is needed to optimize the robot and enhance its effectiveness.