Research on network attack path modeling and risk quantitative assessment algorithm for electric power UAV inspection system

To address the complex cybersecurity threats faced by power inspection unmanned aerial vehicle (UAV) systems in intelligent and networked environments, this paper proposes a method for modeling attack paths and quantifying risks that integrates spatial, temporal, and system-level dimensions. First, a 3D attack path model is constructed by integrating Geographic Information System (GIS) data, mission cycle phases, and hierarchical system vulnerabilities. This model depicts potential attack chains from initial intrusion to system-level disruption, using conditional probability chain rules to quantify path success probabilities. Second, a dynamic risk quantification algorithm based on AHP-fuzzy comprehensive evaluation is designed. It dynamically adjusts risk indicator weights using real-time environmental data to achieve a comprehensive assessment of attack likelihood and impact severity. Experiments conducted in the Gazebo simulation environment based on real transmission line inspection scenarios demonstrate that the proposed method generates highly accurate multi-layer attack paths and effectively identifies critical threat chains. The dynamic risk assessment algorithm significantly outperforms traditional static methods in terms of false positive rate, false negative rate, and trend alignment, exhibiting superior environmental adaptability and practical early warning capabilities. This research provides practical modeling and evaluation tools for securing power UAV inspection systems, demonstrating strong potential for widespread adoption.