The impact of basic traffic flow characteristics on traffic accident occurrence on 2-lane rural roads in Serbia
Year: 2025
Type: article
Abstract: Basic traffic flow characteristics, such as volume and speed, represent the key criteria for estimating the level of service and traffic safety. Numerous studies have been conducted with the aim of determining the impact of the basic traffic flow characteristics on traffic accident occurrence. However, the mutual impact of traffic volume and speed characteristics has not been examined to a sufficient degree. Therefore, the authors of the article analysed the mutual impact of average annual daily traffic (AADT), share of heavy vehicles in the traffic flow (\%HV) and speed parameters (such as speed variance $\left( \sigma^2_{s} \right)$, average travel speed (ATS), speed limit credibility and percentage of exceeding the speed limit (\%ESL) on traffic accident occurrence on Class I 2-lane rural roads in the Serbia. The article analysed spatial distribution and traffic accidents′ severity on homogeneous segments equipped with automatic traffic counters, which provided data on speeds and other traffic flow characteristics during a 5-year period. The application of generalised linear models led to the development of 3 negative binomial models with an natural logarithm (ln) link function for the (1) total number of traffic accidents (ACT), (2) property damage only accidents (ACPDO) and (3) injury and fatality accidents (ACI\&F). The obtained results show that 3 variables are statistically significant in all 3 final models: (1) AADT, (2) $\sigma^2_{s}$ and (3) difference between the free-flow speed and speed limit $\left( \small\Delta S_{ff - {lim}} \right)$. All variables included in the final models are positively associated with the dependent variable. In other words, the rise of the values of AADT, $\sigma^2_{s}$ and $\small\Delta S_{ff - {lim}}$ variables increases the expected number of traffic accidents. The variable related to the \%HV is statistically significant in the models related to the ACT and traffic accidents with injuries and fatalities, and it is positively associated with the expected traffic accident number.
Source: Transport
Institution: University of Belgrade
Language: English
Topic: Traffic and Road Safety
Subfield: Safety, Risk, Reliability and Quality
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $431
OpenAlex ID: W4411097798