Pothole Detection and Warning System for Indian Roads

This paper is based on an application of mobile sensing: sensing and gathering the surface condition of roads. We will fabricate a system and mention the required algorithms to sense the road anomalies by making a portable sensor that can be equipped in any car or public transport. We will call this system pothole detection system (PDS), it will use the mobility of the particular vehicle on which the system will be fitted, and side by side gather data from the vibrations and the GPS sensors, and further process and filter the data to monitor road surface condition. At first, we will deploy the PDS on our own vehicle and test it out in a particular sector of Noida. Using the machine learning approach, we were able to identify and classify the potholes and other road anomalies from the accelerometer data. From the continuous testing and gathering data on a particular stretch of road, we were able to put an algorithm that will successfully detect a pothole with 4.3% chance of failure or if the pothole is too small to be detected. It was further conducted a manual inspection of the reported potholes and found that 80% of the road anomalies reported are in need of serious repair.