Machine Learning-based Smart Helmet for Air Quality and Harmful Event Finding for the Mining Industry

The protection of the mines below ground must be enhanced since catastrophes in subterranean mines are extremely major concerns these days. The problems experienced by miners operating underground include gas explosion, temperatures, etc. If any tragedy happens in mine or any miner is wounded, all the guilt immediately falls on supervisor. So there must be interaction among miners, supervisors and control station. Therefore, the goal of the proposed system is to alter a current mining safety. Helmet represents a piece of the safety equipment miner should wear when mining. The idea is for making the helmet more secure by adding network. This added system is used to sense the circumstances around the miner employed underground and all of real-time values are wirelessly current on the internet through the IoT so that control stations get to know the specific environmental circumstances in which miner employed and if any unusual events occur they are able to offer rescue as quickly as possible. The device also contains the LCD and siren to let fellow workers know when any unpleasant occurrence happens with miner. The suggested system employs multiple sensors as Gas Sensor, DHT11 Sensors, accelerometer, vibrations and IR Sensor. Here the IR sensors is employed as helmet removal sensor. And additionally in this system we utilize machine learning to analyze the employee health, Cardiovascular disease is one of the primary ailments impacting the people throughout the globe. There are various risk factors which contribute to heart disease. The combined use of logistic regression modeling with neural network gives a revolutionary strategy for forecasting the heart disease. Initially a logistic regression analysis is done to find the primary risk variables for predicting the illness. It generates the significant risk variables that are beneficial for forecasting cardiovascular disease according to statistical p-value. The threat variables those aren’t taking the substantial effect are recognized besides deleted. The resulting important issues are presented as inputs to the neural network’s model.