Inspect Road Quality by Using Anomaly Detection Approach

Road quality can be representative of a country's development status, affect transportation speeds and traveler safety. However, there is no standard or set of rules to protect people from the dangers of damaged roads. This study proposes a method for evaluating road quality, which has been made possible by the recent rapid development of information technology and machine learning algorithms, and the popularity and widespread use of smartphones. The proposed inspection method is a road quality inspection APP which collects raw data from smartphone sensors, including GPS and accelerator sensors. Once the data is collected, the server side of the proposed system runs an anomaly detection algorithm to interpret the recorded oscillating amplitude of a specific section of road. The calculated results are then added to the Google Maps app, and abnormal road sections are marked in different colors.