Authors:

• T. Subetha, Department of Information Technology, BVRIT HYDERABAD College of Engineering for Women, Hyderabad, India
• Kayal Padmanandam, Department of Information Technology, BVRIT HYDERABAD College of Engineering for Women, Hyderabad, India
• L. Lakshmi, Department of AI/ML, BVRIT HYDERABAD College of Engineering for Women, Hyderabad, India
• S.L. Aruna Rao, Department of Information Technology, BVRIT HYDERABAD College of Engineering for Women, Hyderabad, India

1. Introduction

Autism Spectrum Disorder (ASD) is a neurodevelopmental condition that significantly affects an individual's ability to communicate and interact socially. Traditional therapeutic approaches, such as Applied Behavioral Analysis (ABA) therapy, have been widely recognized for improving social behaviors in individuals with ASD. However, the challenges associated with these methods, such as the need for one-on-one sessions and consistent real-world practice, have prompted researchers to explore technology-driven solutions.

This case study focuses on the development and application of a Virtual Reality (VR)-based training system designed to enhance social-communication skills in students with ASD using ABA therapy principles.

2. Problem Statement

ASD students often struggle with understanding social cues and engaging in appropriate social interactions. Traditional ABA therapy, though effective, can be time-consuming and resource-intensive. There is a growing need for innovative methods that can provide consistent, immersive, and engaging training to help ASD students improve their social skills.

3. Objectives

The primary objective of this study is to develop an interactive VR-based training system that leverages ABA therapy to teach essential social-communication skills to students with ASD. The system aims to:

• Provide an immersive learning environment where students can observe and practice social behaviors.
• Use AI-driven technologies to recognize and analyze student gestures and speech.
• Implement automatic video self-modelling (VSM) to reinforce learned behaviors.

4. Methodology

4.1 System Design

The VR-based system is designed with two core sessions:

Training Session: In this session, students are exposed to various social scenarios within a virtual environment. The system uses ABA therapy techniques to demonstrate appropriate social behaviors, such as greeting others, maintaining eye contact, and understanding facial expressions. The immersive nature of VR allows students to engage with these scenarios in a controlled and safe environment.

Practice Session: Following the training, students participate in a practice session where they are encouraged to apply the skills they have learned. The system uses a multimodal gesture recognition system and deep neural networks to monitor and analyze the students' gestures and speech. This analysis helps determine the accuracy and effectiveness of the students' social interactions.

4.2 AI and VR Integration

Multimodal Gesture Recognition: The system employs a gesture recognition module that captures and interprets students' hand movements, body language, and facial expressions. This module is crucial for assessing how well students mimic the social behaviors demonstrated during the training session.

Deep Neural Networks: Speech recognition is integrated using deep neural networks to analyze verbal responses. This AI-driven approach ensures that both verbal and non-verbal communication skills are accurately assessed and reinforced.

Automatic Video Self-Modelling (VSM): The system compiles successful interactions into personalized videos, which are shown to the students. VSM allows students to observe themselves performing social behaviors correctly, reinforcing the learned skills and boosting their confidence.

4.3 Comparative Study

A comparative study was conducted to evaluate the effectiveness of the VR-based system. Two groups of ASD students were selected: one group received traditional ABA therapy, while the other group used the VR-based training system. The study assessed improvements in social-communication skills over a specified period.

5. Results

The study revealed that students who used the VR-based training system showed significant improvements in social-communication skills compared to those who received traditional ABA therapy. These students demonstrated better understanding and application of social cues, more consistent verbal and non-verbal communication, and increased confidence in real-world social interactions.

Parents and families of the students reported noticeable positive changes in their children's behavior, highlighting the potential of VR and AI technologies in enhancing therapeutic outcomes for individuals with ASD.

6. Conclusion

The integration of VR and AI technologies into ABA therapy represents a promising advancement in the treatment of ASD. This study demonstrates that a VR-based training system can effectively teach essential social-communication skills to ASD students, offering a scalable and engaging alternative to traditional therapy methods. The positive outcomes observed in this case study suggest that further research and development in this area could lead to even more effective therapeutic tools for individuals with ASD.

7. Future Work

Future research could explore the scalability of the VR-based system to accommodate a wider range of social scenarios and more diverse student needs. Additionally, incorporating real-time feedback and adaptive learning algorithms could further enhance the effectiveness of the training system. Collaborative efforts between educators, therapists, and technologists will be essential in driving the continued evolution of technology-assisted therapies for ASD.