Name
Abstract | ||
|---|---|---|
Cycling helps reduce traffic congestion, environmental pollution and promote a healthy lifestyle for the general public. However, it could also expose cyclists to dangerous environments, resulting in severe consequences and even death. Transport authorities are seeing growing accidents in city regions with increasing cycling population, requiring the development of new risk informed cycling safety policies. This paper aims to develop a new conceptual risk analysis approach based on a Bayesian network (BN) technique to enable the analysis and prediction of the severity of cycling accidents. To identify the risk factors influencing accident severity, 2,269 cycling accident reports from the UK city region were manually collected, where primary data was extracted and analysed. An advanced data training method (i.e. Tree Augmented Naive Bayes (TAN)) for BN development was applied to investigate their correlation and their individual and combined contributions to cycling accident severity. As a result, the risk factors influencing accident severity are prioritised in terms of their risk contribution. The risk levels of accident severity can be predicted in dynamic situations based on the data from simulated and/or real cycling environments. The findings can provide useful insights for making rational cycling safety policies in proportion to different risk levels. |
Year | DOI | Venue |
|---|---|---|
2021 | 10.1016/j.ress.2021.107460 | RELIABILITY ENGINEERING & SYSTEM SAFETY |
Keywords | DocType | Volume |
Cycling safety, Bayesian network, Accident severity, Transport risk analysis | Journal | 209 |
ISSN | Citations | PageRank |
0951-8320 | 0 | 0.34 |
References | Authors | |
0 | 4 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Zai-Li Yang | 1 | 112 | 13.72 |
| Zhisen Yang | 2 | 0 | 1.01 |
| John Smith | 3 | 0 | 0.34 |
| Bostock Adam Peter Robert | 4 | 0 | 0.34 |