Road marking visibility for automated vehicles: Machine detectability and maintenance standards

This study examines the detectability of road markings by automated vehicle sensors across various environmental conditions, focusing on retroreflectivity (R L ) and daytime visibility (Q d ). Controlled tests were conducted on different road marking designs, evaluated with cameras and LiDAR sensors in dry, wet, day, and night conditions. Results show that the slope (m) of all the linear models was positive and statistically significant (p-value < 0.05), confirming that these properties are important for sensor based detection. Current road marking maintenance guidelines primarily focus on R L and Q d values, but often overlook the importance of contrast between the markings and the road surface. This study proposes a method for developing new guidelines that integrate contrast-based criteria and consider several environmental conditions as well as sensor capabilities, to enhance road safety for automated systems without imposing unnecessary challenges on road maintenance standards. The framework promotes consistent road marking visibility, efficient maintenance practices, and supports the safe integration of automated vehicles into transport systems, contributing to the alignment of Operational Design Domains (ODD) with real-world Operational Domains (OD).