Work

Efficient control of drones communications in IoT

Year: 2024
Type: dissertation
Abstract: This doctoral thesis explores the development of efficient drone control methods in the dynamic landscape of drone networks within the Internet of Things (IoT). As drones become increasingly integrated into the IoT ecosystem, addressing the complexities and challenges inherent in their coordination becomes paramount for ensuring reliability and efficiency. The thesis starts with a thorough exploration of IoT concepts alongside drone networks, outlining key application domains and describing state-of-the-art solutions, particularly in localisation and tracking. Additionally, it examines advanced drone route planning strategies, highlighting the opportunities they present and the critical challenges they entail. The main body of the thesis introduces novel cooperative algorithms, drawing from deterministic principles and artificial intelligence (AI) techniques. Inspired by natural phenomena such as flocking birds, these algorithms enable drones to collaboratively determine their routes for tracking mobile sensors within dynamic IoT environments. As the efficacy of these methods is demonstrated, it becomes apparent how they enhance drone cooperation and significantly improve tracking efficiency. Building upon this foundation, the thesis next introduces an innovative deep reinforcement learning (DRL) scheme, empowering autonomous drone agents to efficiently develop optimal data collection strategies within an IoT network. By harnessing DRL, drones continually acquire insights from their environment and actions, adapting to changes and making intelligent decisions to optimise their data collection policies. The scheme adapts state-of-the-art algorithms to effectively scale to high-dimensional state-action spaces commonly encountered in real-world IoT applications. This research contributes to the ongoing discourse surrounding drone-IoT integration, offering novel approaches to drone control. The introduction of these methods opens up new avenues for creating more efficient, and autonomous drone networks within the IoT paradigm, highlighting the untapped potential of AI in this context, and setting the stage for future development in the field.
Language: English
Open Access status: green
OpenAlex ID: W7065324897
Cites: 108
Cited by: 1
Related to: 10