<![CDATA[The Addis Ababa–Djibouti corridor serves as Ethiopia's economic lifeline, handling approximately 95% of the nation's international trade. Despite its strategic importance, the corridor faces significant safety and operational challenges, with heavy goods vehicles involved in 6,914 crashes (14.85% of the total) in 2024, resulting in 715 fatalities. Railway operations are constrained by outdated signaling systems, limiting capacity to only two trains between stations. This study develops an integrated framework combining wireless sensor networks (WSNs) and machine learning for proactive safety monitoring and operational risk management along the corridor. The methodology employs a multi-tier WSN architecture with edge and cloud computing, utilizing gravitational search algorithms for optimal node placement. Machine learning models, Random Forest, XGBoost, and LightGBM, were trained on 2,000 accident records (2018–2023), with SMOTE addressing class imbalance. Random Forest achieved 82% accuracy (AUC-ROC: 0.87), identifying driver age (mean 44 years), nighttime on unlit roads (+62% fatal probability), and rainy conditions as key severity predictors. The proposed WSN-based railway signaling enables 3–4 trains to operate between stations simultaneously, representing a 50–100% capacity increase. Integration of sensor networks and ML enables a fundamental shift from reactive to proactive risk management, with estimated benefits including 15–25% accident reduction, 20–30% maintenance cost savings, and 15–25% extension of infrastructure lifespan. The framework directly aligns with the World Bank’s USD 500 million, Regional Economic Corridor Project and offers a scalable, replicable model for deploying intelligent transport systems across developing economies. This research contributes the first integrated sensor-ML framework for an African trade corridor, offering practical evidence for policy formulation and technology investment prioritization.]]>
Copyrights © 2026
