An earthquake alert system based on a collaborative approach using smart devices

Recently, mobile devices such as smartphones and IoT devices have been successfully used for detecting earthquakes using their built-in accelerometers. To utilize such devices as a seismic sensor, there are two approaches including a standalone or a client-server approach. The client-server approach is more accurate than the stand-alone approach but requires high-performance servers and network infrastructures to process acceleration data captured from many client devices. In the standalone approach, straightforward earthquake detection algorithms can be easily implemented on less powerful mobile devices, but there is a possibility of false alarms. To address this limitation, in this paper, we present a collaborative approach that detects an earthquake using multiple smartphones located in a near area so as to improve the earthquake detection ability of the standalone approach without system and network infrastructures. In our approach, smartphones located in a near area construct a seismic network to detect earthquakes together and then wait for any shakings caused by human activities, mechanical vibrations, earthquakes, etc. When a smartphone detects an earthquakelike motion using an earthquake detection algorithm based on a simple neural network, it broadcasts the detection result to nearby smartphones in a multi-hop manner. Upon receipt of the detection results from nearby smartphones, each smartphone participating in the seismic network performs a decision-making process and confirms an earthquake and then instantiates an alert. The experimental results show the effectiveness of the proposed approach, so we believe that our approach can be effectively used in high seismic countries that lack an earthquake early warning system (EEWS).