Experimental Implementation of COVID-19 Safety Measures in Ride-Sharing Cabs Using Deep Learning and Internet of Things

Abstract

Increased urbanization and on-demand mobility have resulted in the boon of many ride-sharing companies. Besides providing faster, economical, and comfortable rides, these rideshares are also environment-friendly as they save a lot of energy. This research work presents a model which would be beneficial for the passengers riding these carpools as it would not only help to curb the spread of infection in current pandemic but also detect whether the driver is drowsy or not to prevent possible road accidents. The proposed web application includes three detections based on novel deep learning algorithms implementing face recognition, facemask, and drowsiness detection of the driver with an alert mechanism to send immediate email alerts to the company and driver. The novelty of the proposed application is that the current and live status of the driver is continuously recorded using latest technologies like convolutional neural networks (CNNs), histogram of oriented gradients (HOG), support vector machine (SVM) classifier, and computer vision. In addition to this, a real-time vehicle tracking device is also implemented using Node MCU, Global Positioning System (GPS) module, and Blynk app to keep the company updated about the real-time location of the rideshare. The name of the recognized driver is displayed as output. Face mask and drowsiness detection is done with an accuracy of 99%, and the real-time location of the cab is indicated in Google Maps on the Blynk app. The proposed web application would be very beneficial for ride-sharing companies in the current COVID situation.