Protected, or separated, bike lanes have become increasingly common around the United States. Studies have consistently found that people prefer bike facilities that are separated from traffic, such as off-street paths and protected bike lanes, with physical separation such as a post or curb providing increased comfort. The preference for these separated facilities appears to be greater amongst cyclists who ride primarily for recreation (as opposed to for transportation) and among those who cycle less, as well as among the subset of potential bicyclists who are classified as interested in cycling for transportation but concerned about safety and other issues. These results suggest that providing comfortable designs may be vital to expanding the bicycling population beyond current riders. However, these studies of bicyclists’ sense of safety and comfort have generally focused on segments, rather than intersection locations. This report, Contextual Guidance at Intersections for Protected Bicycle Lanes, summarizes a research effort that used a combination of in-person surveys to define user comfort and microsimulations to estimate expected bicyclist and turning-vehicle interactions to estimate bicyclist comfort based on design type and volumes. Findings suggest that protected intersections and bike signals provide the best expected rider comfort, rated by two-thirds of all respondents rated them as very comfortable or somewhat comfortable.

Background: Cities around the United States are increasingly seeking to modernize and enhance their bicycling infrastructure with the aim of safely accommodating an increasing number of cyclists and attracting new cyclists. As a key component of this effort, cities are employing protected (separated) bike lanes. Generally, protected bike lanes assign bicyclists and motorists their own space on the roadway, with some type of vertical separation defining the respective lanes. Intersections pose a challenge; the separation often ends and people riding often must move through areas with cross traffic or turning traffic. Design approaches for protected bike-lane intersections have focused on two main philosophies: 1) reduce separation prior to the intersection by channeling bicyclists toward motor vehicle traffic, and 2) maintain separation at intersections. This research assesses a selection of design approaches for comfort and expected interactions with motorists based on bicycle and turning-vehicle volumes.

Methodology: The research approach is guided by the assumption that cyclist comfort is a key desired design outcome. In-person video surveys are used to identify people’s comfort levels while bicycling through a variety of intersection designs under defined conditions (e.g., with or without interactions with turning motorists). Video data and microsimulation models were used to inform the comparison of the design options and analyze anticipated interactions at various bicycle and vehicle volumes for each of the design options. A total of 277 respondents rated 26 video clips showing cyclists riding through a variety of intersections, for a total of 7,166 ratings. Surveys were conducted at four locations in three states, including urban and suburban locations in Oregon, Minnesota and Maryland. Simulation models were built and calibrated to the designs of bend-in, mixing zones, and bend-out (protected intersection) that were tested in the in-person survey. Bicycle and vehicle volumes varied from 50 to 250 vehicles per hour and 10 simulations pairs were run. Surrogate safety measures were extracted from the resulting 900 trajectory files using the Federal Highway Administration (FHWA) Surrogate Safety Assessment Model (SSAM) software. Comfort estimated from surveys are combined with simulated conflicts for various turning volumes to estimate the level of comfort for the designs.

Findings: Survey ratings demonstrated that designs that minimize interactions with motorists, such as fully separated signal phases and protected intersections, are rated as most comfortable (two-thirds of respondents rated them as very comfortable or somewhat comfortable). Comfort drops off significantly for other designs, particularly for people who are dissuaded from riding due to concerns about traffic (e.g., the Interested but Concerned cyclists). Designs with longer sections of exposure for bicyclists (e.g., via mix or merge areas, or long intersection crossings) were associated with decreased comfort. In general, interactions with turning vehicles caused people to rate cycling comfort lower. Non-mixing zone locations (other than protected intersections) such as bend-in and similar designs were most susceptible to eroded comfort when interacting with turning vehicles. Women and non-white respondents were generally less likely to feel comfortable than other respondents. For the bend-in, mixing zones and bend-out (protected intersection) designs, the research team simulated the expected frequency with which cyclists would encounter turning motorists as a function of through bicycle and right-turn volumes. The research identified exposure distance, measured as the end of vertical separation on one side of the intersection to the start of separation on the far side is a significant predictor of comfort.
The simulation models were calibrated to existing conditions but could not completely represent the interactions between motorists and bicycles. In general, however, as either the bicycle or right-turning vehicle volume increases, the number of simulated conflicts of all three intersection models increases. The number of conflicts per bicycle also increases as the right-turning vehicle volume increases. The simulations found that the number of interactions were highest at the bend-out design, though occurred at lower speeds (the speeds of the turning vehicles were calibrated to each location tested). Due to the number of assumptions required, it is not recommended that the results from the microsimulation be extended outside the context used in this research for weighting the comfort scores.

Guidance: Contextual guidance on selecting intersection design treatments based on estimated cyclist comfort was developed by combining the survey comfort ratings with the simulated frequency with which cyclists would encounter turning motorists. This guidance is provided for both more experienced and tolerant to traffic stress (Bike Inclined) and those less tolerant to traffic stress and sensitive to comfort (Interested but Concerned). The estimated order of comfort for both groups, from least to most comfortable, was mixing zones, lateral shifts, bend-in, and maintain separation, signal and protected intersections. Importantly, comfort scores for the Interested but Concerned groups suggest only the bicycle signal phase separation (3.7 comfort score out of 5 and 65% comfortable) and protected intersection (3.7-3.8 score out of 5 and 67% to 70% comfortable) as recommended designs.