Authors - Rama Satish K V, Hemavathi J, Shrihari Kulkarni, V.V Sathya Sai Likhith, Balasubramani A, Shashaank Raj CS Abstract - The integration of AI-driven voice-controlled systems is transforming human-drone interaction, enabling intuitive and efficient UAV operation. This paper presents a novel approach for voice-commanded drone navigation using a fine-tuned T5 model, ensuring structured execution through predefined flight functions. The system incorporates Whisper for speech-to-text conversion and AirSim for high-fidelity simulation via its Python API. By leveraging NLP and real-time processing, the model accurately maps user commands to drone actions while maintaining execution safety. Experimental results demonstrate 97% accuracy in function mapping, highlighting the system’s reliability and responsive-ness. This work advances AI-powered UAV control, enabling applications in search-and-rescue, autonomous inspections, and hands-free drone operations.
