A comprehensive literature review on intrusion detection datasets for UAVs: taxonomy, challenges, future directions, and recommendations

Unmanned Aerial Vehicles (UAVs) are playing an increasingly important role in several applications, such as agriculture, transportation, and entertainment, raising concerns about their exposure to cyberthreats. Recently, there has been a notable surge in the frequency and sophistication of cyberattacks. In response, the scientific community has leveraged Machine Learning to detect attacks against drones using Intrusion Detection Systems (IDS); however, challenges remain in finding and creating reliable datasets specifically tailored to UAVs. In this study, we investigate UAV-relevant IDS datasets by reviewing articles published between 2019 and 2025. Following PRISMA guidelines, 110 articles were selected for comprehensive review. A key finding was the identification of 53 UAV-relevant IDS datasets, along with their characteristics, limitations, and experimental aspects. This Systematic Literature Review (SLR) introduces a novel taxonomy of datasets and critically compares their categories. Additionally, it provides insights into trends in dataset creation and usage over the past six years. The study discusses associated challenges, such as the absence of standards and legal restrictions, and proposes future directions to address them. Through this SLR, we advocate for sharing UAV-relevant IDS datasets by proposing our paradigm (Plan-Do-Check-Act then Share) and providing a set of 10 recommendations as guidelines to enhance dataset reliability.