Macro-Level Collision Prediction Models Related to Bicycle Use

Abstract

The growing automobile transport results in severe traffic congestion, pollution and road safety problems. Bicycling, as one of sustainable transportation mode, is encouraged in most developed countries for its attributes of convenience, low cost, non-fuel use, and zero-emissions. It is generally accepted that increasing bicycle use could improve road safety. Based on a comprehensive literature review, this paper discusses potential factors influencing bicycle use and bicycle collisions. Understanding these bicycle-related factors is useful to develop new bicycle-related Collision Prediction Models (CPMs) with generalized linear regression. These CPMs can support economic justification of much-needed major bicycle infrastructure investments, and also help policy makers to promote bicycling in an effective and economic manner. Also, a brief methodology of developing such macro-level CPMs is suggested. Based on a case study of City of Kelowna, BC, Canada, several new macro-level CPMs are proposed. Results reveal that the increase of bicycle use can lead to a decrease in total collisions despite an increase in bicycle collisions, which is consistent with the actual case. Also, the bicyclerelated exposure variable, bicycle lane length, has a significantly positive relationship with dependent variables: total collision frequency. In this case, it is concluded that increasing bike lanes (on-street and off-street) can be a good measure to improve road safety. Still, aimed on the research gap, this paper identifies potential works of high interests in future.