Modeling and Evaluating the Safety Effectiveness of Mini-Roundabouts
Mini-roundabouts are a type of roundabout characterized by a small diameter, and fully traversable central island and splitter islands. They are an alternative intersection design option in areas with constraints requiring additional land acquisition. They may be retrofitted within the existing intersection boundaries. They are suited to environments where speeds are relatively low and environmental constraints preclude the use of larger roundabouts with raised central islands. The standard-size roundabouts are safer than traditional minor road stop-controlled or signalized intersections, better suited for traffic calming, and reduce delay as well as emissions. However, the safety benefits associated with mini-roundabouts are not well documented and must be evaluated for planners and engineers to consider more mini-roundabout installations in the United States. Therefore, the focus of this research is on evaluating the safety effectiveness of mini-roundabouts converted from prior control types like two-way stop-controlled or one-way stop-controlled (TWSC or OWSC) and all-way stop-controlled (AWSC) with at least one approach having a speed limit equal to or greater than 35 mph (~56.33 kmph). The methodology includes: 1) identification of mini-roundabout installations in the United States, 2) before and after crash data and traffic volume data collection at selected mini-roundabout locations, 3) before and after analysis for determining safety benefits of mini-roundabouts, 4) safety effectiveness and crash modification factors (CMFs) computation for mini-roundabouts based on before and after crash data, and, 5) examining the effect of traffic characteristics, geometric characteristics, and on-network and off-network characteristics on mini-roundabout safety effectiveness and after period crashes.To accomplish these objectives, 25 mini-roundabout installations in the United States were identified. They are in Georgia (5), Iowa (1), Michigan (4), Minnesota (3), Missouri (1), North Carolina (2), Virginia (1), and Washington State (8). Data pertaining to mini-roundabout geometry, traffic crashes, and traffic volumes were collected from various sources like departments of transportation (DOTs), police departments, Highway Safety Information System (HSIS), Highway Performance Monitoring System (HPMS) database, and state public record centers. At least one year of after period data was available for each selected mini-roundabout.The safety benefits of a mini-roundabout were assessed using naïve before-after analysis employing crashes per year and crash rate as metrics. In naïve before-after analysis, crashes per year in the before period are compared to crashes per year in the after period. The percentage change in the number of crashes per year in the after period from the before period indicates the safety effectiveness of mini-roundabouts. Likewise, the percentage change in the crash rate in the after period was compared with the crash rate in the before period. The safety effectiveness of mini-roundabouts were separately evaluated based on the number of total crashes, fatal and injury (FI) crashes, and property damage only (PDO) crashes. The analysis was carried out separately by prior control types such as two-way stop-controlled (TWSC) or one-way stop-controlled (OWSC) and all-way stop-controlled (AWSC) intersections. The results indicate a decrease in the total number of crashes and the number of FI crashes per year as well as crash rate when a TWSC or OWSC intersection was converted to a mini-roundabout. However, the results indicate an increase in the number of PDO crashes per year while the crash rate remained nearly the same. Similarly, the results indicate an increase in the number of total crashes, FI crashes, and PDO crashes per year and crash rate when an AWSC intersection was converted to a mini-roundabout. The naive before-after analysis based on crashes per year does not account for the effect of exposure (change in traffic volume or other patterns on a selected facility), trend effect (change in traffic composition, driver composition, etc.), and the random effect (regression-to-the-mean bias). On the other hand, before-after crash rate comparison accounts for exposure by considering traffic volume. However, it assumes a linear relationship between crash frequency and traffic volume.Evaluating safety effectiveness using more statistical rigorous techniques such as the Empirical Bayes (EB) method would help in computing a better estimate of safety effectiveness and standard error. Crash and traffic volume data collected for an additional 723 reference intersections were used for safety performance function (mathematical model) development and calibration, and the EB method was used to evaluate safety effectiveness.The safety effectiveness from EB method was computed considering HSM SPFs (calibrated and non-calibrated) and jurisdiction-specific SPFs (calibrated and non-calibrated). The results from the EB method indicate a decrease in the number of total crashes and FI crashes when TWSC/OWSC intersections were converted to mini-roundabouts. However, the results from the EB method indicate an increase in the number of PDO crashes when TWSC/OWSC intersections were converted to mini-roundabouts. The results from the EB method indicate an increase in the number of total crashes, FI crashes, and PDO crashes when AWSC intersections were converted to mini-roundabouts.The safety effectiveness from EB method differed when HSM SPFs and jurisdiction-specific SPFs were used. It also differed when jurisdiction-specific SPFs were used and calibrated for subsequent years. Difference between the safety effectiveness estimates was statistically significant at a 95% confidence level for total crashes at AWSC intersections converted to mini-roundabouts. Further, it also differed when jurisdiction-specific SPFs were developed and compared using 3, 5, 7 and 9 years of crash data. Based on the findings, calibration of jurisdiction-specific SPFs is recommended to account for temporal changes in estimating expected number of crashes in the before and after periods.The CMFs from the EB method are recommended based on calibrated HSM SPFs (TWSC and OWSC), and year-wise calibrated jurisdiction-specific SPFs [OWSC (ramp) and AWSC]. A 22.03% and 61.08% reduction in the number of total crashes and FI crashes but a 4.11% increase in the number PDO crashes is expected when a TWSC/OWSC intersection is converted to a mini-roundabout. Likewise, a 201.45%, 96.20%, and 263.68% increase in the number of total crashes, FI crashes, and PDO crashes is expected when an AWSC intersection is converted to a mini-roundabout. The recommended CMFs for converting a TWSC/OWSC intersection to a mini-roundabout are 0.78 for total crashes, 0.39 for FI crashes, and 1.04 for PDO crashes. Likewise, recommended CMFs for converting an AWSC intersection to a mini-roundabout are 3.01 for total crashes, 1.96 for FI crashes, and 3.64 for PDO crashes. The EB method results indicate that the installation of mini-roundabouts was found to be effective in the reduction of total crashes at 60% of the selected sites (9 out of 15) when TWSC/OWSC intersections are converted to mini-roundabouts They are found to be more effective in the reduction of FI crashes at 90% of the selected sites (14 out of 15). However, they are found to be less effective in the reduction of PDO crashes - at less than 50% of the sites (7 out of 15). Likewise, the installation of mini-roundabouts was found to be effective at only 10% of the selected sites (1 out of 10) for total, FI and PDO crashes when AWSC intersections are converted to mini-roundabouts.Overall, converting a TWSC/OWSC intersection to a mini-roundabout could result in better safety benefits than converting an AWSC intersection to a mini-roundabout. The odds ratio is lower for TWSC/OWSC intersections with a high crash history. However, FI-based odds ratio is higher for mini-roundabouts with a greater number of crashes in the after period. The odds ratio for the number of total crashes and PDO crashes is lower if entry width is higher at AWSC intersections converted to mini-roundabouts. The number of crashes in the before period, cross-street traffic volume, speed limit at major street and cross-street, and intersection skewness have a statistically significant influence on the safety effectiveness of mini-roundabouts (number of crashes in the after period) at a 90% confidence level.