Earthquake Detection and Alerting system using ANN and IOT

Abstract

Earthquakes are one of the deadliest natural disasters. The damage caused by earthquakes is very destructive and the number of deaths are at a very large scale. The National Earthquake Information Centre now detects about 20,000 earthquakes around the globe each year, or approximately 55 per day. Many of the Earth’s earthquake zones coincide with areas of high population density. When a large earthquake occurs in such areas, the results can be catastrophic. The basic parameter to determine an earthquake before it was going to occur is the p wave or primary shock wave which is not destructive to infrastructure because it travels parallel to the propagation of seismic wave. A series of destructive waves that follows p waves are s waves or secondary waves which has a speed of about 60% that of p waves. The movement of the s wave is perpendicular to the direction of propagation of the seismic wave and hence it is extremely risky to infrastructures and cause a lot of damage. This research aims to build an earthquake detection and alerting system using a low-cost accelerometer, ESP32 Wi-Fi module, and ANN algorithm. The input earthquake data is taken by the accelerometer and with the help of the ESP32 Wi-Fi module the data is sent to a system that processes the ANN algorithm with Multi layered perceptron trained with stochastic gradient descent learning algorithm and detects whether an earthquake is going to occur or not and alerts users with ESP32 Wi-Fi module. Key Words: Artificial Neural Networks, ESP32, Accelerometer, Earthquake, Multi-layer perceptron, Sigmoid function, Stochastic Gradient Descent.