Level of Service for Bicycle Through Movement at Signalized Intersections Operating Under Heterogeneous Traffic Flow Conditions

• Published in: Transportation in developing economies (Online) Vol. 3; no. 2; pp. 1 - 16
• Main Authors: Beura, Sambit Kumar, Kumar, N. Kiran, Bhuyan, Prasanta Kumar
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.10.2017; Springer Nature B.V
• Subjects: Bicycles; Civil Engineering; Clustering; Correlation analysis; Developing countries; Development Economics; Engineering; Landscape/Regional and Urban Planning; LDCs; Mathematical models; Original Article; Quality of service; Regression analysis; Sensitivity analysis; Traffic flow; Traffic intersections; Traffic models; Traffic volume; Transportation; User satisfaction; Signalized intersection; Heterogeneous traffic flow; Step-wise regression; Affinity propagation clustering; Sensitivity analysis; Bicycle level of service
• ISSN: 2199-9287; 2199-9295
• DOI: 10.1007/s40890-017-0051-z

This study has taken an attempt to model the quality of services offered to bicycle through movement at urban signalized intersections carrying heterogeneous traffic. An extensive set of quantitative attributes (geometric, traffic and built-environmental) are collected from 35 intersection approaches of four Indian mid-sized cities and are thoroughly analyzed from a bicyclist’s perspective. Perceived satisfactions of approximately 160 bicycle users on each intersection approach are assessed through an extensive perception survey. By using the survey outcomes as the dependent variable, six attributes of intersection approaches having significant influences on the bicycle service quality are identified with the help of Pearson’s correlation analysis. Further, a step-wise regression analysis is carried out to establish a strong relationship between these influencing variables and perceived satisfactions of bicyclists. As observed, the resulting model is highly reliable in the present context and has a very high coefficient of determination ( R 2 ) value of 0.86 with averaged observations. A sensitivity analysis is carried out to recognize the percentage contributions of modeled variables in service quality prediction. As per the results obtained, traffic volume (38%), width (17.6%), and pavement condition (15.9%) of an intersection approach have by far the most significant influences on its service quality. Affinity Propagation (AP) clustering is used to define the ranges of bicycle service categories (A–F). It is observed that approximately 69% of all studied intersection approaches are offering above average quality of services (A–C) and remaining are offering below average (D–F) kind of services at their present scenario.