Building structural analysis based Internet of Things network assisted earthquake detection

Analyzing the patterns of earthquakes and creating a system to detect them, can help to minimize loss of lives. Existing sensor based detection works failed to distinguish between earthquakes and other external vibrational noises that may be caused due to construction or other external sources. In the proposed work, to overcome this drawback, first in-depth structural analysis of shear walls is done to find a structurally suitable location to install the sensor in a building. Further, MPU6050 sensor is used to reduce the complexity and make the circuit less bulky. This sensor consists of an accelerometer which measures the change in acceleration and a gyroscope which records the shaking pattern of the ground. Since this shaking pattern is exhibited exclusively by an earthquake occurrence, the proposed model distinguishes between earthquake and other non-hazardous vibrations efficiently. With the considered model for machine learning algorithms, we were able to achieve 99.82% of accuracy, which is an important factor for earthquake detection.