Multifaceted sensor-based approach for road quality assessment in the Indian road scenario

Abstract

This study aims to explore the feasibility of conducting supervised classification of road barriers in a practical context through the utilization of diverse data collection methods. These methods encompass accelerometer, ultrasonic, GPS, and real-time clock sensors, which collectively contribute to a comprehensive analysis of the subject matter. This study primarily focuses on the highways of India, along with urban and semi-urban areas, as its central subject of investigation. In order to facilitate the collection of data from these sensors, a mobile application referred to as DC has been meticulously developed. In this study, the data collected from the sensors undergo a transformation process to create a fuzzy dataset. This is achieved through the application of min-max normalization followed by fuzzification techniques. A variety of methodologies for measuring distance have been established, each aimed at achieving optimal classification outcomes. One of the primary objectives is to establish comprehensive standards for assessing the condition of roadways, considering a multitude of factors, including the overall length of the road and the extent of any damage present. This study conducts a comprehensive comparative analysis of all distance metrics employed for the classification of road impediments. The findings reveal promising results regarding accuracy, demonstrating an approximate range between 98% and 99%. Furthermore, to facilitate the observation of outcomes in real time, a visualization tool is currently under development. This tool aims to display road obstructions on maps, enhancing the user's ability to navigate and understand the current traffic conditions effectively.