Low-volume roads (LVRs) are an integral part of the rural transportation network providing access to remote rural areas and facilitating the movement of goods from farms to markets. These roads pose unique challenges ...Low-volume roads (LVRs) are an integral part of the rural transportation network providing access to remote rural areas and facilitating the movement of goods from farms to markets. These roads pose unique challenges for highway agencies including those related to safety management on the highway network. Specifically, traditional network screening methods using crash history can be effective in screening rural highways with higher traffic volumes and more frequent crashes. However, these traditional methods are often ineffective in screening LVR networks due to low traffic volumes and the sporadic nature of crash occurrence. Further, many of the LVRs are owned and operated by local agencies that may lack access to detailed crash, traffic and roadway data and the technical expertise within their staff. Therefore, there is a need for more efficient and practical network screening approaches to facilitate safety management programs on these roads. This study proposes one such approach which utilizes a heuristic scoring scheme in assessing the level of risk/safety for the purpose of network screening. The proposed scheme is developed based on the principles of US Highway Safety Manual (HSM) analysis procedures for rural highways and the fundamentals in safety science. The primary application of the proposed scheme is for ranking sites in network screening applications or for comparing multiple improvement alternatives at a specific site. The proposed approach does not require access to detailed databases, technical expertise, or exact information, making it an invaluable tool for small agencies and local governments (e.g. counties, townships, tribal governments, etc.).展开更多
The conditional generative adversarial network(CGAN)is used in this paper for empirical Bayes(EB)analysis of road crash hotspots.EB is a well-known method for estimating the expected crash frequency of sites(e.g.road ...The conditional generative adversarial network(CGAN)is used in this paper for empirical Bayes(EB)analysis of road crash hotspots.EB is a well-known method for estimating the expected crash frequency of sites(e.g.road segments,intersections)and then prioritising these sites to identify a subset of high priority sites(e.g.hotspots)for additional safety audits/improvements.In contrast to the conventional EB approach,which employs a statis tical model such as the negative binomial model(NB-EB)to model crash frequency data,the recently developed CGAN-EB approach uses a conditional generative adversarial net work,a form of deep neural network,that can model any form of distributions of the crash frequency data.Previous research has shown that the CGAN-EB performs as well as or bet ter than NB-EB,however that work considered only a small range of crash data character istics and did not examine the spatial and temporal transferability.In this paper a series of simulation experiments are devised and carried out to assess the CGAN-EB performance across a wide range of conditions and compares it to the NB-EB.The simulation results show that CGAN-EB performs as well as NB-EB when conditions favor the NB-EB model(i.e.data conform to the assumptions of the NB model)and outperforms NB-EB in experi ments reflecting conditions frequently encountered in practice(i.e.low sample mean crash rates,and when crash frequency does not follow a log-linear relationship with covariates).Also,temporal and spatial transferability of both approaches were evaluated using field data and both CGAN-EB and NB-EB approaches were found to have similar performance.展开更多
文摘Low-volume roads (LVRs) are an integral part of the rural transportation network providing access to remote rural areas and facilitating the movement of goods from farms to markets. These roads pose unique challenges for highway agencies including those related to safety management on the highway network. Specifically, traditional network screening methods using crash history can be effective in screening rural highways with higher traffic volumes and more frequent crashes. However, these traditional methods are often ineffective in screening LVR networks due to low traffic volumes and the sporadic nature of crash occurrence. Further, many of the LVRs are owned and operated by local agencies that may lack access to detailed crash, traffic and roadway data and the technical expertise within their staff. Therefore, there is a need for more efficient and practical network screening approaches to facilitate safety management programs on these roads. This study proposes one such approach which utilizes a heuristic scoring scheme in assessing the level of risk/safety for the purpose of network screening. The proposed scheme is developed based on the principles of US Highway Safety Manual (HSM) analysis procedures for rural highways and the fundamentals in safety science. The primary application of the proposed scheme is for ranking sites in network screening applications or for comparing multiple improvement alternatives at a specific site. The proposed approach does not require access to detailed databases, technical expertise, or exact information, making it an invaluable tool for small agencies and local governments (e.g. counties, townships, tribal governments, etc.).
文摘The conditional generative adversarial network(CGAN)is used in this paper for empirical Bayes(EB)analysis of road crash hotspots.EB is a well-known method for estimating the expected crash frequency of sites(e.g.road segments,intersections)and then prioritising these sites to identify a subset of high priority sites(e.g.hotspots)for additional safety audits/improvements.In contrast to the conventional EB approach,which employs a statis tical model such as the negative binomial model(NB-EB)to model crash frequency data,the recently developed CGAN-EB approach uses a conditional generative adversarial net work,a form of deep neural network,that can model any form of distributions of the crash frequency data.Previous research has shown that the CGAN-EB performs as well as or bet ter than NB-EB,however that work considered only a small range of crash data character istics and did not examine the spatial and temporal transferability.In this paper a series of simulation experiments are devised and carried out to assess the CGAN-EB performance across a wide range of conditions and compares it to the NB-EB.The simulation results show that CGAN-EB performs as well as NB-EB when conditions favor the NB-EB model(i.e.data conform to the assumptions of the NB model)and outperforms NB-EB in experi ments reflecting conditions frequently encountered in practice(i.e.low sample mean crash rates,and when crash frequency does not follow a log-linear relationship with covariates).Also,temporal and spatial transferability of both approaches were evaluated using field data and both CGAN-EB and NB-EB approaches were found to have similar performance.
