Quality of Bicycle Traffic Management at Urban Road Links and Signalized Intersections Operating under Mixed Traffic Conditions

This study proposes bicycle level of service (BLOS) models for the urban road links and signalized intersection approaches operating under heterogeneous traffic flow conditions prevailing in developing countries. Extensive data collection was carried out through field investigations, videography techniques, and perception surveys to inspect which parameters primarily affect the quality of bicycling in the mentioned facilities. As many as 74 road links and 70 intersection approaches of nine different Indian cities were included in the study locations. In addition, the perceived satisfaction ratings of 22,320 on-site bicyclists (i.e., 155 bicyclists from each site) were included in the database. These ratings ranged from 1 to 6 (excellent-worst bicycling conditions), and were used as the array of dependent variable in the model development process. Separate BLOS models were developed for road links and bicycle through movement at signalized intersections through the application of a novel evolutionary artificial intelligence technique, namely multi-gene genetic programming. The link model included eight significant variables in total whereas the intersection model included a total of seven. With compacted structures, both models represented excellent reliabilities in the present context and produced high coefficient of determination (R-2) values of above 0.86 with averaged observations. These models will help traffic planners and engineers to evaluate the service levels of urban bicycle facilities of developing countries. Statistical analysis of modeled parameters has reported that the traffic volume and crossing pedestrians have the highest influences on the quality of bicycle traffic management at urban road links and intersection approaches, respectively.