LIVES: XAI-based Secure UAV Communication Framework for Essential Items Delivery

The application of Unmanned Aerial Vehicles (UAVs) for securely delivering essential items during extreme conditions includes numerous challenges, particularly in ensuring identification of potential threats. To solve this challenge, we propose a strong framework that makes use of Boosting algorithms, with a particular focus on XGBoost to enhance attack classification in UAV-based delivery systems. The XGBoost model achieved an accuracy of 93%, a precision of 0.9312, a recall of 0.9311, and an F1 score of 0.9283 , which makes it a highly suitable model for security of UAVs. Explainable AI (XAI) is used to further enhance classification including SHAP and LIME to identify the most influential features, such as the barometer and battery levels. Our framework provides a comprehensive solution for secure UAV-based deliveries. The experimental results proves the effectiveness of the proposed model which shows its potential for real-life applications.