National Cooperative Highway Research Program Project 3-70, Multi-modal Level of Service Analysis for Urban Streets, was developed in response to a need expressed by state and local departments of transportation to analyze transportation facilities from a multi-modal and user perspective. The Transportation Equity Act for the 21st Century (TEA-21) called for the mainstreaming of transit, pedestrian and bicycle projects into the planning, design and operation of transportation systems, however, tools and methods do not currently exist that allow engineers and planners to examine the tradeoffs associated with improvements to one mode or another. This research paper describes the modeling efforts towards the development of an auto LOS model that reflects the perceptions of drivers on urban streets. A data collection approach using video laboratories and a modeling approach suitable for the discrete and ordered nature of the response variable were selected. The approach enabled the estimation of the distribution of LOS responses given a set of explanatory variables that describe the geometry and operational effectiveness of the urban street facility. The analysis indicated that the participants responses tended to be highly variables, and that models that characterize the distribution would be preferable than point estimates. A model that uses the number of stops per mile experienced on the facility, and the presence of exclusive left turn lanes at the intersections gave the best fit to the data, and matched the mean clip LOS rating 71% of the time. It is clear that significant uncertainty remains in the prediction of LOS, and it is likely that additional factors beyond the narrow HCM purview may need to be added to improve on future models explanatory power.
Abstract