A Framework for Enhancing Open and Distance Learning Using Large Language Models and Augmented Reality

Abstract

Open and Distance Learning (ODL) increasingly demands intelligent, interactive, and adaptive platforms to bridge the gap between conceptual understanding and learner engagement. This paper presents ARPhysicsLab, a framework that integrates Large Language Model (LLM) and Augmented Reality (AR) to enhance immersive learning in physics education. The system employs LlaMA-3.1-8B as the core reasoning engine, enabling context-aware dialogue, feedback, and content generation. A Python Flask middleware serves as the communication bridge between the LLM backend and the Unitybased AR module, ensuring efficient real-time synchronisation and modular scalability. The framework follows a Design Science Research approach encompassing requirement analysis, prototype development, and user evaluation. A pilot study involving 30 ODL learners assessed engagement, usability, and perceived learning improvement using a 5-point Likert scale. Results indicate high learner engagement (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$M=4.6$</tex>), strong usability ratings (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$M=4.4$</tex>), and a 21 % post-test conceptual gain, confirming the framework's pedagogical value. The proposed framework demonstrates the potential of integrating LLMs with AR through lightweight middleware to achieve adaptive, scalable, and transformative ODL experiences.