Unmanned Aerial Vehicle (UAV) Control Through Speech Recognition

This paper delves into the performance of a voice-activated Unmanned Aerial Vehicle (UAV) control system in terms of speech recognition. Through an evaluation encompassing accuracy, confidence levels, and overall performance, we scrutinize the system’s efficacy utilizing live microphone input under varying noise conditions (ideal and moderate) as well as pre-recorded audio files. Our findings elucidate the system’s proficiency in executing complex commands within noise-free environments, juxtaposed with deteriorated performance in the presence of background noise and for basic commands. While pre-recorded audio attains near-perfect accuracy, its controlled setting fails to mirror real-world challenges adequately. This study underscores the inherent trade-off between command complexity and noise tolerance, highlighting the system’s challenges with basic keyword recognition and the imperative need for enhanced noise cancellation algorithms to ensure robust real-world operation. We advocate for future research endeavors to concentrate on these areas to bolster the reliability and user-friendliness of voice-activated UAV control systems.