Traffic injury prevention · 2026
Hybrid modeling of unobserved heterogeneity in Ohio freeway crash severity using RWIS data: A Random Forest-Correlated Mixed Logit approach.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Bukuru Esther, Balyagati Philip F, Sando Thobias, Kidando Emmanuel, Owusu-Danquah Josiah
Аннотация
OBJECTIVE: Weather-related crashes continue to contribute disproportionately to severe injuries and fatalities on high-speed freeways despite advances in Intelligent Transportation Systems (ITS). This study evaluates the safety impacts of real-time weather conditions on freeway crash injury severity across Ohio using integrated Road Weather Information System (RWIS), crash, and roadway data.
METHODS: Hourly RWIS observations, crash records, and roadway characteristics from 2019-2023 were integrated, covering approximately 452,292 freeway crashes. A hybrid analytical framework combining Random Forest variable selection and Correlated Mixed Logit with Heterogeneity in Means (CMXL-HM) modeling was applied to capture unobserved heterogeneity in driver responses. An hourly exposure-based Severity Index (SI) was developed to normalize crash frequencies across weather states and quantify relative severity risk under varying environmental conditions.
RESULTS: Severity Index results indicated that freezing temperatures and moderate wind speeds exhibited the highest relative severity risk. The CMXL-HM model significantly outperformed baseline Multinomial Logit and Mixed Logit models based on likelihood ratio tests (χ2 = 21,754.69, p < 0.001) and goodness-of-fit measures. Freezing temperatures (≤32 °F) and wind speed (5-10 mph) were identified as random parameters with statistically significant heterogeneity, while other parameters exhibited more consistent effects. Several adverse weather conditions exhibited protective effects after controlling traffic and roadway factors, suggesting risk-compensating driver behavior. Marginal effects showed that higher traffic volumes and rear-end collisions increased crash severity, while morning periods and the spring season reduced it. A strong and statistically significant positive association (ρ = 0.937) between freezing temperatures and travel speed indicates compounded high-risk conditions.
CONCLUSIONS: Findings indicate that real-time weather conditions exert heterogeneous and nonlinear effects on freeway crash severity. The proposed hourly exposure-normalized framework provides actionable insights for identifying high-risk weather thresholds and supports evidence-based deployment of RWIS infrastructure, weather-responsive Variable Speed Limits, and targeted traveler information systems to enhance freeway safety management.
METHODS: Hourly RWIS observations, crash records, and roadway characteristics from 2019-2023 were integrated, covering approximately 452,292 freeway crashes. A hybrid analytical framework combining Random Forest variable selection and Correlated Mixed Logit with Heterogeneity in Means (CMXL-HM) modeling was applied to capture unobserved heterogeneity in driver responses. An hourly exposure-based Severity Index (SI) was developed to normalize crash frequencies across weather states and quantify relative severity risk under varying environmental conditions.
RESULTS: Severity Index results indicated that freezing temperatures and moderate wind speeds exhibited the highest relative severity risk. The CMXL-HM model significantly outperformed baseline Multinomial Logit and Mixed Logit models based on likelihood ratio tests (χ2 = 21,754.69, p < 0.001) and goodness-of-fit measures. Freezing temperatures (≤32 °F) and wind speed (5-10 mph) were identified as random parameters with statistically significant heterogeneity, while other parameters exhibited more consistent effects. Several adverse weather conditions exhibited protective effects after controlling traffic and roadway factors, suggesting risk-compensating driver behavior. Marginal effects showed that higher traffic volumes and rear-end collisions increased crash severity, while morning periods and the spring season reduced it. A strong and statistically significant positive association (ρ = 0.937) between freezing temperatures and travel speed indicates compounded high-risk conditions.
CONCLUSIONS: Findings indicate that real-time weather conditions exert heterogeneous and nonlinear effects on freeway crash severity. The proposed hourly exposure-normalized framework provides actionable insights for identifying high-risk weather thresholds and supports evidence-based deployment of RWIS infrastructure, weather-responsive Variable Speed Limits, and targeted traveler information systems to enhance freeway safety management.