Variations in skid resistance measurements over time have been observed since measurements began and have been the subject of many studies aimed at developing a better understanding of these variations. However, there is no clear consensus on the approach to quantify seasonal variation of skid resistance measurements. Nonetheless procedures such as seasonal correction of annual network survey data and long-term monitoring of benchmark sites in England, coupled with QA procedures and standards that define the operational conditions for skid resistance measurement devices enables measurements to be adjusted to account for allows these variations and so minimise their impact on the management of skid resistance.
This report presents a research study that investigated different approaches of assessing short- and long-term climatic impacts on skid resistance variation on the Strategic Road Network (SRN) in England. As part of the project, a machine learning based model for skid resistance relationship with weather conditions was developed. For that purpose, skid resistance data from multiple sideway-force skid resistance measurement devices covering the period 2013-2019 were used. These data were combined with relevant historical texture data and weather data from nearby eather stations. Promising results suggested to extend the model to a selection of benchmark sites with the aim to adapt the model to a wider range of surfaces and locations across the English SRN.
Overall, the developed machine learning model showed potential to predict skid resistance values, but further model improvement would be needed in order to provide more consistently accurate results. The inclusion of traffic information was suggested as a way to improve the model’s accuracy as the skid resistance variation would then be taking account of both changes in weather and traffic conditions.
