Virtual Real-Time Artificial Intelligence Learning Platform for Adaptive Drone Operation Training Using Digital Twin and Vision-Based Simulation

Abstract

The rapid advancement of Unmanned Aerial Vehicles (UAVs) has significantly expanded their applications in surveillance, disaster management, precision agriculture, logistics, infrastructure inspection, and defense. Consequently, the demand for competent drone operators has increased substantially. Conventional drone training primarily relies on physical flight exercises, which require considerable operational costs, specialized instructors, extensive training areas, and involve risks of equipment damage and safety incidents. These limitations highlight the need for a safer, more efficient, and adaptive learning platform. This study proposes a Virtual Real-Time Artificial Intelligence (VRTAI) learning platform that integrates Digital Twin technology, Large Language Models (LLMs), Vision Transformer (ViT), and real-time flight simulation into an intelligent educational environment for drone operation training. The proposed platform provides immersive simulation scenarios, AI-assisted instruction, automatic performance assessment, and adaptive learning based on trainee competency. A Digital Twin replicates the physical UAV and its operational environment, enabling realistic flight dynamics and mission planning. Vision-based AI performs object detection and situational awareness training, while the AI tutor delivers interactive guidance and immediate feedback. The system architecture consists of five major components: a virtual simulation engine, AI learning engine, Digital Twin module, real-time analytics dashboard, and learning management system. Performance evaluation is conducted through simulation scenarios involving manual flight, autonomous navigation, waypoint missions, emergency procedures, and object detection tasks. Learning outcomes are assessed using flight stability, navigation accuracy, collision avoidance, mission completion time, and AI-assisted competency scores. The proposed framework demonstrates that integrating artificial intelligence with virtual real-time simulation can significantly enhance learning effectiveness, reduce operational costs, improve safety, and provide personalized learning experiences. This platform offers a promising solution for future drone education and professional operator certification programs.