Influence of Pavement Marking Angular Systems on Visibility Predictions Using Computer Models

Abstract

Computer models like CARVE (Computer Aided Roadmarking Visibility Evaluator), which predict the nighttime visibility of pavement markings, require quantitative information about the coefficient of retroreflected luminance RL for selected pavement marking material and selected dimension parameters. To specify the coefficient of retroreflected luminance RL, four angles are necessary (observation, entrance, presentation, and orientation). However, RL values for pavement marking materials are measured as a function of two angles only, whereas the other two are held constant. This two-angle–only approach makes the RL data-collection procedure and subsequent computer models significantly less cumbersome. The objective was to determine the angleranges in real-world driving situations for the International Commission on Illumination’s goniometer, intrinsic, application, and roadmarking four-angle system to measure RL values of different pavement marking materials for sets of the observation, entrance, presentation, and orientation angles and to compare these RL values with RL approximations calculated by using two-angle intrinsic and roadmarking angular systems. RL approximations using the two-angle intrinsic system (based on observation and entrance angles only) overestimate the actual RL values by between 4 and 10 percent. On the other hand, the RL approximations based on the two-angle roadmarking system (taking into account only illumination-elevation and the observation-elevation angles) overestimate the actual RL value by between 25 and 82 percent. Under retroreflective conditions, it is suggested to use the two-angle intrinsic system rather than the two-angle roadmarking system for modeling pavement marking visibility problems because the abridged intrinsic system yields more accurate and more conservative RL approximations.