Innovative AI Robot Aims to Support Students with Autism in the Classroom
Imagine a classroom where a friendly robot helps predict and prevent behavioral meltdowns before they happen. This could become a reality thanks to an innovative project led by a team of dedicated students.
The COVID-19 pandemic exacerbated mental health and behavior-related challenges in the classroom, especially for individuals with Autism Spectrum Disorder (ASD), who often struggle with communication and social interactions. These students often require extra attention at a time when there are fewer teachers available, so there’s an urgent need for behavioral interventions.
Deitra Kuester, an associate professor in education and Saifuddin Mahmud, assistant professor in computer science and information systems teamed up to design a multidisciplinary capstone project for computer science students. The goal of the project was to design a kid-friendly robot using artificial intelligence (AI) to help predict behavioral meltdowns.
Four students chose to work on the RASD (Robots for Autism Spectrum Disorder) CatBot based on their personal experiences and interactions with individuals with exceptional learning needs. Team members included Anna Breuker, who graduated in May and served as the team lead, David Henvick, Taylor Sims and Jasmine Gonzalez.
“The project allowed them the chance to not only design and build a machine that could be used by others, it also afforded an opportunity to engage in cross-discipline collaboration about a real-world problem,” Kuester explained. Team members engaged in training and research sessions, visited schools and classrooms and interacted with students with varying learning needs.
“Our group built this robot from the ground up with little to no previous robotics experience, so even small accomplishments like having it move for the first time or programming the face that's displayed was incredibly rewarding,” Breuker said.
The biggest challenges, according to Sims, were learning the robot operating system (ROS) and constructing the robot. “It took weeks to gather the right parts and integrate them with ROS and there were times we couldn't progress due to missing parts or unavailable professors. None of us had prior experience with electrical engineering or robotics, making the process particularly challenging,” she said.
The CatBot, designed to be approachable with its cat-like appearance, uses advanced AI to monitor facial features, sounds, and movements, alerting teachers to potential meltdowns before disruptions occur. Students presented the RASD CatBot at the Student Expo last spring and received positive feedback from attendees.
Henvick, who describes himself as being on the high end of the spectrum, had a personal stake in this project. “Personally, I'm fairly high functioning, but if there was something like this available when I was younger, it could have helped me pick up on social cues faster.”
This RASD CatBot project will continue this fall with a new group of students. Breuker said, “I won't be there to see what happens, but I really believe it could become something truly helpful if more people get involved and improve the baseline that we worked so hard to achieve.”
—Emily Potts
