E-way approach to sense the road condition using machine learning

Abstract

Smartphones are mainly functional to be adopted as a money-spinning and easy to execute tool for the
measurement of road surface roughness condition, which is very essential for road monitoring and maintenance
planning. In this study, an experiment has been carried out to collect data from accelerometers and gyroscopes on
smartphones. The collected data is processed in the frequency domain to calculate magnitudes of the vibration. Road
roughness condition that is modeled as a linear function of the vibration magnitudes, taking into account of both data
from accelerometer and gyroscope as well as the average speed, achieves better estimation than the model that takes into
account the magnitude from the accelerometer and the average speed alone. The finding is potentially significant for the
development of a more accurate model and a better smartphone app to estimate road roughness condition from
smartphone sensors.