ROBOT-BASED THERAPEUTIC INTERVENTION FOR CHILDREN WITH AUTISM
(Project funded by Fundamental Research Grant Scheme (FRGS) from the Ministry of Higher Education Malaysia. “Human-Robot Interaction (HRI) Algorithm in Robot-Based Intervention of Children with Autism Spectrum Disorders (ASD)“)
In recent years, noticeable progress has been made in using robotics for people challenged with Autism Spectrum Disorder (ASD), in areas like training social skills, promoting explorative behavior and early diagnosis. ASD is a lifelong developmental disability that effects the brain’s normal development of social and communication skills. Since autism has no cure, an early, intensive and appropriate therapy and treatment program will greatly improve the outlook for most young children with autism. Although there is growing evidence that people with ASD respond positively to the robot interventions, each person has its own specific challenges, which requires therapy tailored to the individual. Although therapists are open for adopting robotics and other technology platforms, they require systems that they can use without high technical skills (like programming), and are specific and flexible to suit different needs. Motivated by these emerging needs, the main objective of this research is to investigate and model a basic robotic platform that can be used by therapists to easily create and edit exercises for children with ASD through a humanoid robot. A Human-Robot Interaction (HRI) concept is selected as the base algorithm since it has great potential to teach children with autism how to detect and understand emotions and social behaviors. In fact the biggest challenge faced by children with autism is the difficulty in communicating and understanding social interactions.
The methodology and outcome of this research are outlined in three main stages. First is to determine behavior analysis and comparison data of normal children and children with ASD when they interact with humanoid robot that programmed with basic HRI architecture. The outcome of this stage is a set of qualitative and quantitative analysis data consists of behavior pattern, self-motivation, emotion characteristics, and improvement of physical action. Based on these data, the second stage is to program a new algorithm based on evolutionary HRI architecture called Architecture for Close Human-Robot Integration (ACHRIN) in humanoid robot platform consists of standard ASD assessments. The algorithm combines robot-assisted intervention concepts of developmental, individual-difference and relation-based requirements. The outcome of this stage is an algorithm that facilitates a non-expert person to be involved in the robotics operation when needed. The final stage is to evaluate the proposed algorithm in therapy model involving real autism children and actual occupational therapist.
This research is carried out in collaboration with the Faculty of Medicine, UiTM, the Rehabilitation Center, Sungai Buloh Hospital, and the National Autism Society of Malaysia (NASOM). This project is expected to deliver a basic HRI-based platform tailored with individual needs of autism community, and for further exploration on rehabilitation and therapy robotics that will benefit many people not only in Malaysia, but also in other countries.
Recent Progress and Outcome
We have conducted initial behavior response experiment with 12 autistic children using humanoid robot NAO that programmed with therapeutic robot-based intervention modules. Evaluation was conducted using Behaviour Score Sheet that was designed based on GARS-2. The results from three subscale; stereotype behavior, communication, and social interaction shows significant decrease of autism characteristics of the autistic children. Comparison was made with normal classroom session.