| Abstract | This project explores the development of AI-powered Virtual Reality (VR) and Augmented Reality
(AR) systems to teach social skills to individuals with Autism Spectrum Disorder (ASD), utilizing the
Proximal Policy Optimization (PPO) algorithm for reinforcement learning. Social communication
challenges are central to ASD, and traditional interventions may not offer the flexibility or scalability
needed for individualized learning. By leveraging PPO, a state-of-the-art reinforcement learning
algorithm, the system dynamically adapts to the user’s behavior, learning style, and progress to create
personalized social scenarios. The PPO algorithm is used to optimize the system’s behavior, ensuring
it can adjust the difficulty and complexity of social interactions based on real-time feedback from the
user. The AI agent, embedded within the VR/AR environment, interacts with the user by presenting
socially relevant tasks—such as maintaining eye contact, understanding facial expressions, and
engaging in group conversations. As the user interacts with the virtual environment, the PPO algorithm adjusts the agent’s responses, maximizing positive outcomes such as appropriate social responses and minimizing undesirable behaviors like social withdrawal or misunderstanding. By using PPO, the system allows for continuous learning and improvement. It generates increasingly complex scenarios while considering the user’s previous actions, providing timely and accurate feedback. This enhances the ability to master social skills in a safe, immersive setting, leading to more effective interventions for individuals with ASD. The ultimate goal is to improve social skills, build confidence, and enable greater independence for individuals with ASD in real-world social interactions.
|
|---|
