Current haul road management techniques, such as routine, periodic and urgent maintenance have shortcomings in
many complex haul road environments. Real-time road condition monitoring may significantly reduce maintenance
costs, both to the road and to the vehicles. A recent idea is that vehicle on-board data collection systems could be
used to monitor haul roads on a real-time basis by means of vibration signature analysis. This paper proposes a
methodology based on Bayesian regression to isolate the effect of varying vehicle speed on the measured vehicle
response metric. A key feature of the proposed methodology is that it avoids the costly need to generate analytical
or empirical vehicle models.