Robot-Assisted Diagnosis of Autism Spectrum Disorder
The diagnostic assessments for Autism Spectrum Disorder (ASD) are complex and require specialized expertise to integrate information from various sources, such as parent reports and clinical observations, in order to make an accurate diagnosis. Unfortunately, diagnostic facilities for ASD are often limited and located in suburban areas, making them inaccessible to a large portion of the population.
Embracing technology to unlock the hidden potential and diagnose autism, paving the way for a brighter future.
To address this issue, we propose harnessing the inclination of children with autism to interact with technological tools by utilizing Socially Assistive Robots (SAR) for diagnosing ASD in children. Our approach involves creating a framework for multi-modal behavior analysis, focusing on the analysis of speech, facial expressions, and gestures of children with autism. During a diagnostic session, the child's behavior is recorded, and the audio-video data is carefully examined to identify characteristic behaviors that can aid in the diagnosis of autism. Leveraging the capabilities of SAR and utilizing advanced analysis techniques, we aim to improve the early diagnosis of autism by identifying specific behavioral patterns that are indicative of the disorder.
We focus on addressing the following broader research questions:
- How do children of Indian ethnicity, both with and without autism, respond to robot-assisted interventions?
- In what ways can Robot-Assisted Systems (RAS) supported by multimodal behavior analysis be deployed as a supportive tool to assist experts in the diagnosis of Autism Spectrum Disorder (ASD) in children?
Students : B. Ashwini (IIIT-Delhi)
Superviser : Dr Jainendra Shukla (IIIT-Delhi)
Funding: Supported by Start-up Research Grant of Science and Engineering Research Board, Government of India (Ref: SRG/2020/002454)