Earthquake Prediction by Generating and Detecting Seismic Wave

This project aims to develop a real-time earthquake detection and classification system using a combination of hardware, machine learning, cloud computing, and mobile application development. The system utilizes an MPU6050 accelerometer and vibration sensor to continuously monitor seismic activity, with data transmission handled by an ESP8266 WiFi module to store readings in Firebase for further analysis. The collected data is labeled under various conditions, namely normal, vibration, and earthquake, to train a Random Forest classifier that could distinguish between the seismic states effectively. The model is then used to classify the real-time inputs from the sensor, thus correctly detecting the events of earthquakes. To make the system more user-friendly, a Flutter-based mobile application is developed for real-time visualization of sensor data and state classification. The system sends out instant voice notifications and phone vibrations whenever it detects an earthquake-level vibration. The model was tested using different levels of vibration by simulating various seismic conditions, and the results proved that it was reliable in the detection of earthquakes. While the current implementation performs well, future enhancements could include integrating additional sensors, refining the classification model, and optimizing real-time data processing for increased accuracy. This project presents a scalable and practical earthquake detection system, leveraging sensor technology, machine learning, and mobile applications to provide an efficient early warning mechanism for earthquake-prone regions.