Drone Security in Connected Agriculture: Threats, Datasets, and AI-Driven Solutions
Anis Charfi, Samiha Ayed, Lamia Chaari Fourati · 2025
This positioning paper analyzes the security vulnerabilities in drone-based agricultural systems through the lens of threats, datasets, and AI-based solutions. The commercial drone industry, expected to expand from a value of ${\$}$38.25 billion in 2023 to ${\$}$244.95 billion in 2032 [7], makes the need for security increasingly pressing. We make four primary contributions: (1) a systematic classification of cyber threats to agricultural drone systems, segmented by target layers; (2) an assessment of current security datasets, from an agriculture-centric perspective; (3) a new categorization framework of AI-based security solutions by operational phase and technology category; and (4) identification of research gaps, particularly in reaction-phase and proximal generative AI solutions. This method also offers actionable insights into dataset selection for training purposes, attack comprehension, and AI solution design, all aimed at enhancing the security of agricultural drones.