Secure Disaster Management with Blockchain-based Mutual Authentication and Machine Learning for Drought Prediction

Current Disaster Management Systems (DMS) often struggle to effectively mitigate threats due to inherent limitations in both data security and real-time analysis capabilities. While the Internet of Things (IoT) provides valuable real-time data for prediction, it also introduces security vulnerabilities that can undermine system integrity. One significant threat, drought, remains particularly challenging as existing solutions fail to offer reliable real-time detection mechanisms. To address these gaps, this work proposes a novel blockchain-based framework that harnesses the power of IoT, Machine Learning (ML), and the InterPlanetary File System (IPFS) to enhance DMS effectiveness across all phases. Our solution bridges a critical gap by enabling real-time drought prediction through secure IoT communication, fortified with hyperelliptic curve cryptography, and scalable data storage facilitated by IPFS. Additionally, the integration of ML algorithms, specifically CNN and XGBoost, achieves drought prediction accuracies of 70.6% and 72.1%, respectively, showcasing a significant improvement in predictive performance.