As cycling continues to grow in popularity, the volume of bicycle traffic is increasing in many metropolitan areas. To accommodate this trend, a shift from supply-oriented to demand-oriented design approaches will be necessary. This shift requires an empirical foundation on which to develop design guidelines with recommended bicycle facility widths that vary depending on the expected bicycle traffic demand. This paper presents an analysis of cross-sectional data collected over 5 months using inductive loop detectors in Muenster, Germany. The data included the timestamp accurate to 1 s, speed, direction, and inductive loop ID for over 1.3 million bicycle crossings. We analyzed the relationships between speed, density, and flow across the full width of the bicycle facilities, as well as separately within the right-hand, center, and left-hand sublanes. Our findings confirmed a decrease in average speed with increasing flow across the entire path. However, the sublane-specific analyses indicated important differences: the right-hand sublane exhibited the lowest average speed and the most pronounced decrease in speed with increasing volume, whereas a relatively stable speed was maintained on the left-hand sublane. Under free-flow conditions, density increased linearly with flow across the full cross-section. This increase was strongest in the right-hand sublane, followed by the center, and was minimal in the left-hand sublane, reflecting cyclists’ tendency to keep to the right. We present an analysis of how density propagates laterally from the right-hand side across the bicycle facility and propose a method for determining the necessary width of unidirectional bicycle facilities.
Kaths et al. (Wed,) studied this question.