Research on the optimization method of a complex diagrammatic sign on urban expressways based on the non-integer rank RSR method

Abstract

To investigate the optimal method of the complex diagrammatic guide signs (DGSs), a typical complex DGS is selected, and five alternatives are considered, including the current situation (CS), repetition (RT), simplification (SF), pavement words (PW), and advance placement (AP). This study conducts a driving simulation experiment and develops a comprehensive index system based on five aspects: operating status, maneuvering behavior, lane change behavior, subjective perception, and errors. Seventeen indicators are extracted and analyzed in total. The repeated-measure analysis of variances is performed for the overall, and segment-by-segment influence. In the overall analysis results, the significance indicators are primarily operating status, lane change behavior, subjective perception, and errors. The gas pedal and the releasing gas pedal distance were also significantly affected. However, braking-related indicators are not greatly influenced. In the segment-by-segment analysis results, the five operational status indicators, gas pedals and lane numbers are mainly significantly impacted. It also obtains a spatial distribution pattern of the significance indicators, whose location is related to the area of DGS settings in various alternatives. Significant differences exist between the overall analysis and the segment-by-segment analysis. Significant impact indicators are selected based on two types of analysis. The non-integer rank RSR method is employed to evaluate the efficiency of five alternatives. The final rank order from best to worst was RT, AP, CS, PW, and SF. Comparatively, drivers in RT and AP will experience more minor speed fluctuations, less driving time and throttle release distance, earlier early lane change behavior, and lower error rates. This study recommends the RT and AP alternatives to improve the complex DGS. Under specific conditions, the AP option is preferred.