Post-Earthquake Network Restoration: Statistical Seismic Road Closure Prediction and Efficient MDRU Routing

Abstract

Post-earthquake scenarios have brought connectivity challenges to the forefront of research in recent years. Particularly, the randomness and large-scale of road and telecom network infrastructure damage within the aftermath hinders communications coverage restoration during the most critical hours when lives are at stake. This paper proposes a seismic-based post-earthquake city and cellular network model to statistically predict the status of road closures and base station failures based on fundamental earthquake measurements. The presented model considers a generic Manhattan grid-based city model, with buildings featuring random heights. In addition, it quantifies the probability of building collapse and the consequent probability of road closure which accounts for the random debris nature. Moreover, the model accounts for the dependencies between the debris width, height, and the relative location with respect to the earthquake epicenter. Furthermore, a routing algorithm for movable and deployable resource units (MDRUs) that exploits the derived statistical model is proposed to ensure that MDRUs are efficiently deployed and connectivity is restored swiftly. The proposed routing algorithm is extensively tested over a large set of simulation scenarios depicting different earthquake magnitudes and was shown to provide up to 31% traveling time reduction compared to a blind distance-based approach. Finally, the conducted simulations showed the effectiveness of the proposed MDRUs deployment approach in restoring the communications coverage from a signal-to-interference plus noise ratio perspective in the majority of the considered locations.