42. Road Surface Quality Monitoring Using Machine Learning Algorithm

Nowadays, analyzing the road surface conditions is one of the most important aspects of road infrastructure which in turn leads to the better driving conditions and minimizes the risk of road accident. Traditional road condition monitoring systems falls short of collecting real-time update about the road conditions. In earlier models, road surface condition monitoring is done for the fixed roads and static speed of the vehicles. Various systems have proposed approaches of utilizing the sensors mounted in the vehicles. But this approach will not help in predicting the exact location of the potholes, speed breakers and staggered roads. Therefore, smartphone-based road condition assessment as well as the use of the navigation has gained a great existence. We propose to analyze different machine learning approaches to effectively classify the road conditions using accelerometer, gyroscope and GPS data collected from smartphones. In order to avoid noise in the data, we also captured the videos of the roads. This dual technique to data collection will help in providing a more accurate location of potholes, speed breakers and staggered roads. This way of data collection using machine learning algorithms will help in the classifications of roads conditions into various features such as smooth roads, potholes, speed breakers and staggered roads. This information will be provided to the user through the map by classifying the various road conditions. Accelerometers and Gyroscope sensors will investigate various features from all the three axis of the sensors in order to provide a more accurate location of classified roads. Investigate the performance using SVM, random forest, neural network and deep neural network to classify the road conditions. Hence, our results will show that the models trained with the help of the dual technique of data collection will provide the more accurate results. By using neural networks will provide significantly more accurate data classification. The approaches discussed here can be implemented on a larger scale to monitor road for defects that present a safety risk to commuters as well as provide maintenance information to relevant authorities.