Quick decentralized estimation of earthquake epicenter with low-cost IoT network

When an earthquake occurs, it is crucial to quickly estimate its epicenter coordinates and the most affected areas to inform rescue decision-making. Current seismological services perform this estimation relying on centralized processing. On the other hand, we propose a fully decentralized architecture that uses edge computation on a network of sensing and communication low-cost IoT nodes. Nodes exploit M2M messages with neighbors to exchange early warning alerts and to grasp relevant information to estimate the epicenter in real time with a lightweight three-step algorithm. Following an iterative process, the pipeline produces a good estimation with an average error as low as 3km. We describe the system architecture and the proposed strategy and analyze the performances of each estimation step.