Virtual reality (VR) table tennis

Abstract

Virtual Reality (VR) gaming has seen significant advancements in recent years. It provides players immersive experiences that mimic real-world interactions. This project aims to enhance the realism and interactivity of VR table tennis games by developing an AI opponent trained using Machine Learning (ML) algorithm as compared to the convention rule-based opponents. The research objectives include exploring the feasibility of implementing MLagents in VR gaming environments and assessing their effectiveness in creating challenging opponents.This project will be utilizing ML-Agents Toolkit to implement Proximal Policy Optimization on the table tennis paddle which will act as the agent. The project has achieved the basic goal of creating an AI opponent which can return the ball through inference. Future Improvements could involve improvement of ball physics to simulate spinning effects of the ball and training the model with 3-dimensional variables with greater computing resources.Overall, this project has successfully demonstrated the feasibility integrating ML into VR, allowing future game developers to create immersive and adaptive gaming experiences that push the boundaries of VR entertainment.