Road quality is an important issue in everyday life for all car owners. This issue seems to be critically important in underground mines, where LHD machines are used for material transport. One of the biggest problems for LHD operation is relatively quick tires degradation. One of possible reasons might be road surface quality, indeed. However, driver's skills as well as ways of machine operation (loading, acceleration, breaking...) might also play a crucial role. Nowadays, many of machines are equipped with onboard monitoring system that allows to monitor basic parameters (speed, torque, temperatures, pressures etc.) at some predefined components. To complete the picture, we propose to use proposed already (but not for mining applications) vibration measurement for road roughness evaluation. To measure vibration acceleration is relatively easy task (we used simple smartphone here), unfortunately method of parametrization and concluding about road quality is still a challenge in mining case. In this paper we have presented a short communication related to first experimental work and some ideas how to deal with this problem using statistical tools for signal modeling.