PathGuard: Dynamic Large Vehicle Detection and Real-time Alerts on Narrow Roads Using Mobile Sensors

Abstract

On a narrow road, an accident is hard to avoid even for a responsible driver. If vehicles are stuck in traffic, driving on a single lane is worrying and takes time. For small vehicles, narrow roads pose unique challenges, especially in identifying large vehicles, hence reducing the likelihood of an accident. The study discovers these issues and presents how an inventive Intelligent Transportation System (ITS) has been developed as a worldwide phenomenon that aims at enhancing safety on narrow roads by integrating with mobile sensors. Smartphones are used by almost everyone today because their prices have gone down. The study examines the effectiveness of different machine learning models for the task of classifying vehicle type using (accelerometer, and gyroscope) sensors. The results reveal that the Random Forest model is the most effective having a mean accuracy rate of 99.78 %. Moreover, the trained Random Forest Model has been combined with an originally developed unique warning algorithm that integrates geofencing methods for drawing polygons around narrow roads and location data from smartphones. To summarise, this study adds to the development of safety systems in transport and offers useful ideas for developing and implementing real-time safety applications for narrow roads.