Journal of Transportation Engineering, Part A: Systems

Vehicle–pedestrian crashes are one of the common road accident types. Pedestrian crossing warrants are guidelines that recommend a particular type of pedestrian crossing facility to be provided at a crossing location. Various countries around the world have their crossing warrant guidelines, but most of them are focused toward at-grade facilities. Even with existing warrant criteria, the number of pedestrian–vehicle crashes is alarming. This indicates that the warrant criteria and their threshold values may need reexamination. Several of the existing warrant guidelines do not provide the required clarity for the provision of grade-separated pedestrian crossing facilities (GSPCFs). GSPCFs include overpasses (or foot overbridges), underpasses (or subways), and partial underpasses (or hump subways). The existing warrant guidelines consider GSPCFs as a single unit and do not identify which type of GSPCF should be provided at a crossing location. Appropriate GSPCFs not only minimize the accident rate but also eliminate delays for both vehicles and pedestrians. This paper conducts an in-depth analysis of the existing literature on urban midblock GSPCF warrants around the world. The discussion highlights the limitations associated with existing warrant guidelines and the need to formulate improved GSPCF warrants.

This study develops a set of crash modification factors (CMFs) to evaluate the effects of lowering urban-road speed limits on vehicle and pedestrian safety. Cross-sectional methods and observation before–after methods are used to develop CMFs. In general, a CMF estimates the expected change in the frequency of crashes after specific countermeasures are applied on the road. In this study, the safety improvement effect in the section adjacent to the applied section as well as the section for which the policy to lower the speed limit was applied were evaluated. The results indicate that lowering the speed limit is effective in reducing the number of crashes. In particular, the CMFs for crashes involving serious injury and death are 0.6656–0.7804 in the application sections and 0.7979–0.8273 in the adjacent sections. This means that lowering the speed limit can reduce not only the number of crashes but also the occurrence of serious crashes. This study can be used to promote safety by analyzing the effect of the policy to lower the speed limits in the future and can be applied to the evaluation of the effectiveness of various safety policies in cities.

Traditional two-lane or multilane roundabouts gradually become less efficient and safer to operate with increasing traffic volumes due to the weaving area. To mitigate this problem, several new types of roundabouts, including the turbo roundabout, have been gradually invented. Turbo roundabouts have been widely used, attributing to their unique lane separation facilities which eliminate the weaving area and improve the operational efficiency and safety of the roundabouts. However, the most common turbo roundabouts are single-lane or two-lane ones with insufficient capacity. Therefore, this paper focuses on the operational features’ evaluation and forms selection of multilane turbo roundabouts using the entropy weight method (EWM). Based on the commonly utilized operational efficiency indicators, three additional indicators are selected to ensure the results’ accuracy. The EWM is innovatively used to calculate the weights of adopted indicators and achieve a comprehensive evaluation of different roundabouts under various traffic scenarios. Furthermore, the operational features of different roundabouts are simulated and analyzed in Verkehr in Staedten Simulation (VISSIM). The final results demonstrate the EWM’s utility in selecting optimal solutions. The spiral turbo roundabout showed the worst operational efficiency. The conventional form roundabout is proven suitable for small traffic volumes, whereas the rotor turbo roundabout performs better under large traffic volumes. In addition, a modified rotor turbo roundabout is presented, and the final results place its operational efficiency between those of the conventional form and the rotor. The turbo roundabouts achieve the most significant improvement in delays and the number of stops, potentially improving these aspects by more than 50%.

Abstract

Access to transportation is a vital issue for persons with disabilities (PwDs). This paper aims to define measures to improve the mobility, road safety, and access to transportation of drivers with disabilities (DwDs) who use hand controls (HCs) and encourage persons with disabilities to drive. Measures were proposed based on identified factors that influence the mobility with private vehicles and the participation in traffic accidents of drivers with disabilities who use hand controls. The research was conducted in Serbia, a country with a lower inclusion rate of persons with disabilities. The dependent variables in the research were average annual mileage as an indicator of mobility and the number of traffic accidents with adapted vehicles per year as an indicator of road safety. The four most important areas for action with seven measures were recognized. The areas of action are (1) improving the driving training process, (2) appropriate vehicle adaptations, (3) improving awareness about road safety, and (4) supporting persons with disabilities to drive.

Abstract

In this paper, we defined measures to improve the mobility, road safety, and access to transportation of drivers with disabilities who use hand controls and encourage persons with disabilities to drive. Based on a comprehensive analysis, we proposed the following measures: individual approach at the driving training process, training in driving simulators, periodic adjustment of hand control’s performance, encouraging using vehicles with emerging technologies, education about risky behaviors in traffic, developing appropriate applications for travel planning, and more support for specific groups (e.g., females). These measures would significantly improve the quality of life of persons with disabilities and contribute to achieving the United Nations’ Target 11.2, improving the accessibility of transport for persons with disabilities. Because the measures were defined on a sample of respondents in Serbia, a country with a lower inclusion rate, their applicability is possible to other regions with similar problems.

This study investigates a driving behavior referred to as the cooperative safety concept that assumes for strings of conventional vehicles traveling in nighttime or conditions with reduced visibility during a stable flow condition, the leading vehicles would bear much of the navigational challenges and risks while the following vehicles enjoy reduced driving workload and improved navigation safety. The study includes a comprehensive investigation of the safety risk levels and driver behaviors at intersections and freeway ramp locations in an attempt to verify this phenomenon using data from the large-scale Second Strategic Highway Research Program (SHRP 2) naturalistic driving study database. Overall, the driver behavior analysis showed that drivers following other vehicles tended to travel at lower speeds but with more acceleration activities than other vehicles. In addition, lighting during nighttime appeared to help alleviate the behavioral differences between the two types of travelers and resulted in more dispersed merging, diverging, and lane-changing behaviors. The safety event data analysis showed that higher traffic levels tended to correlate with more safety events in general but significantly fewer single-vehicle events. In addition, higher traffic levels correlated with a significantly lower likelihood of crashes in general when a safety event occurred. In the SHRP 2 data, safety events included crashes, near crashes, and statistically selected baseline events recorded during the data collection. The findings of this study, including in particular the event analysis, indicated that vehicles following other vehicles in a free-flow condition tended to drive slower and have lower safety risks in terms of crashes in general and single-vehicle crashes in particular. This knowledge can have significant implications for applications such as advanced lighting systems, cooperative vehicle features, and smart traffic control strategies.

Reliable, real-time prediction of delay and density is challenging as direct measurement of these variables is difficult. Though studies yielding reasonably accurate predictions of delay and density are reported in the literature, a comprehensive methodology to simultaneously predict both delay and density is lacking. Hence, a recursive technique that uses minimal real-time data for dynamic simultaneous prediction of midblock density and intersection delay is proposed. This study uses conservation equation-based recursive prediction of the number of vehicles inside the midblock section (density), which in turn is used to predict delay using shockwave theory. The Kalman Filter is a one-step-ahead density prediction method that can yield reliable density predictions even under the presence of errors in detector data. The one-step-ahead delay predictions obtained had a Mean Absolute Percentile Error (MAPE) of 10.4%, whereas the one-step-ahead density predictions obtained had a MAPE of 9.96%. Due to its robustness, this method can be used to arrive at one-step-ahead predictions of parameters like delay and queue length for any traffic scenario for which shockwave diagrams can be produced.

Several traffic studies necessitate vehicle counting during peak hours and throughout the day, as well as detailed classification and tracking, which consumes human time and effort, particularly at intersections. Manual efforts primarily collect the necessary traffic demand data live in the field or from video recordings using an extensive data manipulation process. Alternative solutions include computer-based systems that perform human tasks more efficiently and with less time and effort, and these systems vary in function and performance. This paper proposes a comprehensive computer-based system that detects, tracks, and computes related statistics in real time while making the best use of available resources, such as public road surveillance cameras. The main contribution of this work is the effectiveness of combining various computer vision algorithms to achieve high-accuracy performance during real-time streaming of road cameras. The experiments confirm the system’s performance by achieving an average success rate of 93.2%. The novel aspect of this work is that detections, point extractions, matching, tracking, and classification were implemented in a single system that ensures real-time execution and high-accuracy output and uses existing infrastructure. The system compensates for the variations in light between day and night and between cloudy and sunny weather. It also recovers hidden vehicles and changes in view for each vehicle as it moves. The proposed method efficiently and partially integrates some mechanisms into a single system.

Many North American jurisdictions are creating walkable urban environments through compact Complete Streets (CS) geometric designs, while not addressing the mobility and accessibility of goods despite the essential role goods movement plays in sustaining the liveability of the community. Prescriptive curb radii limits recommended by CS guidelines to lower pedestrian crossing distances may not adequately accommodate the right-turn maneuver of a truck. The paper develops a performance-based decision support tool to guide the design of urban intersection curb radii that facilitate the safe and efficient accommodation of trucks and pedestrians. The decision support tool relies on a novel Freight-Walkability relationship to define the context of urban intersections and establishes a curb radius design domain. A case study demonstrates the quantification of the Freight-Walkability relationship (in terms of peak hour right-turning truck volumes and a proposed Walkability Index) and the application of the tool at an existing intersection in Winnipeg, Canada. The tool helps transportation engineers and planners balance the mobility needs of trucks and pedestrians through short-term street-level design changes and long-term land use transformations.

A sound barrier, also called a noise barrier, plays an irreplaceable role in traffic noise abatement. The Federal Highway Administration’s (FHWA) highway noise regulation requires each state highway agency to maintain a complete inventory of all constructed noise-abatement features. Although key information for most of the newly constructed sound barriers has been inventoried, public transportation agencies are still struggling to keep close track of the in-service barriers because their inventory information is nonexisting, and manual inventory remains time-consuming, labor-intensive, and often dangerous. Therefore, many agencies have shown more interest in exploring the possibility of using light detection and ranging (LiDAR) data for assisting in sound barrier inventory, thanks to the widely available data set and much-improved data quality. This study proposes a LiDAR-based sound barrier inventory method to automatically extract the sound barrier’s location and measure the corresponding geometry. The extraction of a sound barrier is achieved using its unique features after random sample consensus (RANSAC)-based ground extraction and region-growing segmentation. The geometry measurement of the sound barrier is performed by analyzing the detailed dimension of the extracted point cloud, including location, height, length, and lateral offset. The experimental test conducted near Carver, Massachusetts showed the results with a precision rate of 99.9% and a recall rate of 93.8%. Moreover, the outcome of the experimental test has demonstrated the robustness of the proposed method in different complexities of the background and sound barrier types (linear and zigzag). This study has demonstrated the feasibility of using LiDAR for effectively inventorying in-service barriers. Besides the critical application of asset management, the detailed location and geometry information provided by the proposed method can provide valuable insight for other critical applications, such as traffic noise modeling.

The focus of this paper is on evaluating the safety effectiveness of converting a stop-controlled intersection with a speed $\mathrm{limit}\ge 56.3\mathrm{km}/\mathrm{h}$ ($35\mathrm{mi}/\mathrm{h}$) to a miniroundabout and examining the role of influencing factors on their safety effectiveness in the United States. Crash, traffic volume, and geometry data for 25 miniroundabouts in eight states was collected to conduct before–after analysis using the empirical Bayes (EB) method. In addition, crash and traffic volume data for 723 reference intersections were gathered and used for computing the calibration factors and developing jurisdiction-specific safety performance functions (SPFs). A 19.63% and 60.55% reduction in the number of total crashes and fatal and injury (FI) crashes but an 7.72% increase in property damage only (PDO) crashes was observed when a two-way/one-way stop-controlled (TWSC/OWSC) intersection was converted to a miniroundabout. A 224.76%, 74.30%, and 282.71% increase in the number of total crashes, FI crashes, and PDO crashes was observed when an all-way stop-controlled (AWSC) intersection was converted to a miniroundabout. Converting a TWSC/OWSC intersection to a miniroundabout has better safety benefits than converting an AWSC intersection to a miniroundabout. The recommended crash modification factors (CMFs) for converting a TWSC/OWSC intersection to a miniroundabout are 0.80, 0.39, and 1.08 for total, FI, and PDO crashes. The recommended CMFs for converting an AWSC intersection to a miniroundabout are 3.25, 1.74, and 3.83 for total, FI, and PDO crashes. 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 miniroundabouts at a 90% confidence level. These findings are useful to researchers and practitioners for conducting safety benefit analysis and making informed decisions pertaining to converting a stop-controlled intersection to a miniroundabout.

This research evaluated the effect of the COVID-19 social isolation orders on traffic volume, traffic violations and road crashes in the city of Fortaleza, Brazil. Using data from automated traffic enforcement cameras, a reduction in traffic volume between 30% and 50% was observed during the social isolation period. However, even with the traffic volume reduction, the absolute number of speeding and red-light running violations were 13% and 26% higher than prepandemic levels, respectively. When controlling for traffic exposure, the violation rates increased by more than 100%. After social isolation restrictions were lifted and the traffic volumes returned to prepandemic levels, both traffic violations and traffic violation rates remained at elevated levels (14% to 44% higher than prepandemic levels), possibly related to a nationwide decision that delayed the issuing of violation tickets. Using an interrupted time-series approach and segmented Poisson and negative binomial regression models, it was found that the fatal crash rate was 1.66 times greater during the period of social isolation compared to the prepandemic levels but returned to prepandemic levels following the removal of the social isolation restrictions. A significant reduction in injury crash rate was observed during and following the period of social isolation restrictions; however, the authors hypothesize that this is related to injury crash underreporting during the pandemic.

Emerging mobile Internet applications have become valuable data sources for fine-grained transportation analysis, which allows the introduction of the concept of Personalization in both microscopic and macroscopic modeling of travel behaviors and traffic dynamics. Inspired by personalized recommendation systems, the personalized transportation models emphasize the importance of individual and local information. Two representative cases are presented in this study and two architectures, namely the travel behavior modeling architecture and the geoinformation modeling architecture, are proposed to address the problems of bike-sharing destination prediction and ensemble of ride-hailing demand predictors, respectively. Their performance has been verified by two case studies using the Mobike bike-sharing data and the DiDi ride-hailing demand data.

Abstract

Virtual test driving is a methodology to evaluate vehicle performance and environmental impacts under realistic driving conditions. This approach can generate comprehensive traffic information and trajectory data that reflect the dynamic and stochastic properties of transportation systems. Despite the benefits of virtual test driving, it is challenging to ensure a sufficient correlation between virtual and various real-life traffic situations. In this study, the simulation environment comprises a microscopic traffic network in the district of Aachen and a drivetrain model representing the longitudinal dynamics of a real car. In addressing the research challenge of sufficient correlation between virtual and real traffic, this work aims to develop a methodology to calibrate and validate virtual test driving within a selected study site by comparing simulation results with multi-source field measurements. The comparison between the field data and simulation is threefold: We used traffic counts to evaluate the simulated traffic volume at the macroscopic level; the drone-captured trajectory data to characterize the individual vehicle movements on highways, roundabouts, and intersections; and the real driving data of probe vehicles to investigate the influence of vehicle trajectories on energy demand. We also developed a generic approach to investigate the correlation between driving cycles and drivetrain performance, accounting for different combinations of real and virtual objects in three calibration and validation stages. Furthermore, the concept of mapping jigsaw puzzle scenarios is proposed to replicate the drone-captured concrete scenarios in the simulation and extend the calibrated results to a large-scale network. The results indicate that the proposed toolchain can effectively cope with the validation complexity of virtual test driving. A comparison with additional field measurements reveals the benefits of virtual test driving in providing a wide spectrum of realistic driving conditions in a cost-efficient and time-saving manner.

Practical Applications

Virtual test driving is a methodology to evaluate vehicle performance and environmental impacts under realistic driving conditions. Compared to laboratory tests and field tests, virtual test driving is cost-efficient and time-saving. However, it is challenging to ensure a sufficient correlation between virtual and various real-life traffic situations. This paper proposes a methodology for calibrating and validating virtual test driving with multi-source field data, such as traffic volume at the macroscopic level; drone-captured trajectory data on highways, roundabouts, and intersections; and naturalistic driving data of probe vehicles. In this study, the simulation environment comprises a microscopic traffic network in the district of Aachen and a drivetrain model representing the longitudinal dynamics of a real car. The simulation results were compared with field data based on different measures, such as average speed, acceleration, standstill time, and energy demand. The calibrated and validated simulation environment can reflect the dynamic nature of transportation systems in a stochastic manner.

Abstract

The pandemic of COVID-19 has affected human patterns since December 2019. Since the very beginning, most countries imposed strict measures such as lockdowns and the suspension of all nonessential movements to reduce the spread of the pandemic. Therefore, mobility, road safety, and travel behavior were also significantly affected. At present, many studies tried to investigate travel or mobility behavior changes taking into account all possible transportation modes, but very few studies investigated driving behavior. This study aims to investigate driving behavior and its correlation with the strictness of COVID-19 response measures. Four neural network autoregression (NNAR) models with an external regressor were developed in order to forecast three different future stringency scenarios. NNAR models were employed as the forecasting performance was superior when comparing with statistical autoregressive integrated moving average (ARIMA) models. The NNAR models were developed using driving behavior-related variables (i.e., driving speed, speeding, speeding duration percentage, and mobile use percentage), derived from a smartphone application that has been developed by OSeven Telematics. The NNAR models were trained on 2020 data and three different scenarios were predicted for 2021 by providing three different constant stringency indices (i.e., 0, 55, 85). In particular, normal conditions without restrictions were simulated with zero stringency index, whereas moderate restrictions were simulated with 55 and finally, fully restrictions were simulated with 85. The NNAR modeling results showed that with higher stringency index, mobile use and driving speed tend to increase, whereas speeding duration demonstrates higher peaks. Interestingly, with stricter response measures, lower values were forecasted for speeding. Taking into account the modeling outcomes, there is a direct effect of the COVID-19 response measures on driving behavior. Nevertheless, a wider time frame for data collection as well as the use of more sophisticated techniques to control for the interrelationship between COVID-19 spread and driving behavior might be useful for future studies.

Practical Applications

Interested stakeholders could exploit the study findings and the lessons learned during the pandemic in order to mitigate road safety implications. The COVID-19 pandemic demonstrated the vulnerability of mobility, travel behavior, road safety, and driving behavior patterns throughout health or societal crises. For example, driving speed, peaks of speeding duration percentage, as well as mobile use were found to be higher for an increased stringency index (i.e., during measures imposition) which lower traffic volumes imply. As a direct effect between response measures and driving behavior was observed, measures that promote safety and equity could be tested using the modeling approach followed in this paper. In this direction, road safety administrations have already proposed lowering speed limits inside cities to 20 or $30\mathrm{km}/\mathrm{h}$. Lower speeds could reduce the chance of road crashes, serious injuries, fatalities, and harsh events. Active traveling such as walking, cycling, or e-scooters could also be promoted by policymakers as a step towards safer and environmentally friendly mobility.

The focus of this study was twofold: (1) to estimate models of frequency and injury severity in pedestrian crashes at signalized intersections and (2) to examine whether the “safety in numbers” effect applies to pedestrian safety in the United States. Specifically, the analysis used novel and robust measures of pedestrian exposure: pedestrian crossing volumes estimated from 1 year of pedestrian push-button traffic signal data. Multiple negative binomial models—predicting 10-year pedestrian crash counts at 1,606 signals in Utah—were estimated, accounting for different levels of data availability. The models showed similar results, identifying specific characteristics of the signals that saw more pedestrian crashes. These characteristics were higher volumes of pedestrian and motor vehicle traffic, longer average crossing distances, more crosswalks, continental instead of standard markings, no prohibitions of right-turns-on-red, no bike lanes, near-side instead of far-side bus stops, greater shares of commercial or vacant land uses, more employment density, greater intersection density, no schools or places of worship, and greater shares of people with a disability or people of Hispanic or non-White race/ethnicity. To analyze injury severity in pedestrian crashes, an ordered logit model was fitted with 1,483 pedestrian crash observations. The model results indicated that vehicle size, vehicle maneuvering direction, lighting conditions, and involvement of teenage/older driver and DUI/drowsy/distracted driving in crashes were significantly associated with pedestrian crash severity. Notably, the study also found strong support for the “safety in numbers” effect, in which pedestrian–vehicle crash rates decline with an increase in pedestrian volumes.

This study aims to explore the role of built environments affecting emergency medical services (EMS) responses to traffic crashes. Specifically, this study integrated socioeconomic databases with a crash database that contains the EMS response information. Given the multilevel data structure, a hierarchical model was developed to connect EMS response times to the built environment and other associated factors at various hierarchies. The model results revealed that the built environment plays a vital role in EMS performance in terms of response times. For example, EMS response times differ significantly between rural and urban areas. If other factors are held constant, the EMS response time for a rural crash is 25.13% more likely to be longer than 10 min than for an urban crash. Other factors such as land use, area, development, roadway class, road lighting, weather, and EMS facility distance are also significantly related to EMS response times. This study offers insights into improving EMS responses to traffic crashes by considering the role of built environments. More implications are discussed in the paper.

Uncertainty quantification is important for making reliable decisions in transportation planning and operations. In the field of short-term traffic condition forecasting, uncertainty quantification methods include primarily distribution-based approaches and nondistribution-based approaches. For the former, the generalized autoregressive conditional heteroscedasticity (GARCH) model has been widely applied to model and quantify traffic condition uncertainty in terms of prediction interval under normality assumption. However, this normality assumption has not been systematically investigated yet. Therefore, this paper attempts to investigate this normality assumption and thereby quantify traffic condition uncertainty, using a method with steps of residual calculation and investigation, normality investigation, distribution estimation, uncertainty quantification, and performance measurement. Using real-world traffic flow data, the distributions of the selected samples are shown to be nonnormal using the Kolmogorov-Smirnov test and normal probability plot. Distribution estimation using nonnormal models shows that the $t$ location-scale distribution and generalized error distribution (GED) can be used to model traffic condition uncertainty. Uncertainty quantification using GARCH under these nonnormal distributions further show that nonnormal models outperform the normal model, with the GARCH model under $t$ location-scale distribution yielding the best performance. Future studies are recommended to promote the investigation into traffic condition uncertainty quantification and application.

With the advance of emerging vehicle technologies, more and more work has been devoted to analyzing the resulting impacts in traffic efficiency and energy savings, whereas relatively limited studies have investigated the impact of emerging vehicles on traffic safety. To bridge the research gap, this study selected two emerging vehicle technologies, adaptive cruise control (ACC) of automated vehicles (AVs) and cooperative and adaptive cruise control (CACC) of connected automated vehicles (CAVs), to analyze the impact of emerging vehicles on traffic safety in both homogeneous and heterogeneous traffic. In homogeneous traffic, theoretical linear stability analysis and simulation both indicated that CACC has more robust string stability and is safer compared to ACC due to the fact that CACC is able to transmit acceleration information to the following vehicle. For heterogeneous traffic, microscopic simulations with different vehicle classes were conducted with consideration of downgradation and platooning intensity. Simulation results showed that for mixed CAV and human-driven vehicle (HV) traffic, the collision risk tended to increase first and then fall with the market penetration rate (MPR) of CAVs, which indicates that the moderate deployment stage is the most dangerous stage. In addition, results showed that higher platooning intensity led to lower collision risk in both mixed CAV and HV and AV and HV traffic. As a consequence, this study suggests that vehicle platooning and strengthening the connectivity for emerging vehicles can improve the longitudinal traffic safety of emerging vehicles.

This study explored how the presence of work zones could influence the saturation flow rate (SFR) prevailing at an intersection. Specifically, it researched construction-ridden intersections with interweaving movements (CIWIMs) of vehicle flows that proceed across the stop line and down the connective lanes on the downstream approach to the adjacent intersection. First, image recognition and tracking algorithms were used to extract 2,545 vehicle trajectories from the video captured on-site. Then, the trajectories were manipulated based on lanes to obtain the saturated headway time of the entry-lane stop line during effective green time and the variables related to lane-change behaviors after passing the stop line (e.g., lane-change percentage, lane-change position, lorry percentage, and average passing speed). In addition, certain linear and nonlinear regression methods were employed to estimate lane-focused SFR models in a parsimonious fashion. Subsequently, the Highway Capacity Manual (HCM) model, along with Schroeder’s model, was pairwise compared with the newly proposed Box-Cox model for validation. The results indicate that the mean errors are 28.86% and 17.70% for the HCM and Schroeder’s models, respectively, while the estimation error for the Box-Cox model is merely 7.20%. This sensitivity analysis reveals that the proportion of bidirectional lane changes, spatial use rate of lane changes, and proportion of heavier vehicles significantly compromises the CIWIM-based SFR. One important finding is that the models accounting for microscopic channel-change behaviors, with higher estimation accuracy compared with existing models, can also be used for traffic simulation parameter calibration and road delay estimation to obtain higher validity and precision.

Disruptions due to either natural or anthropogenic hazards can significantly impact the operation of critical infrastructure networks (e.g., transportation systems) as they may instigate network-level (cascade) systemic risks, thus impacting the overall city resilience. Recent relevant studies demonstrated the need to quantify the resilience of city infrastructure networks following failures of one/some of their main components, considering both topological and operational network measures. Subsequently, focusing on robustness (a key resilience attribute) and on transit (a major critical infrastructure network), the current study develops a related quantification tool employing a hybrid approach that integrates complex network theoretic measures with data analytics, and specifically clustering and genetic algorithms. To demonstrate the practical utility of the developed tool, the robustness of the City of Minneapolis bus transit network is quantified under possible cascade failures represented by node (i.e., bus stop), link (i.e., route segment), and route failure scenarios. The robustness quantification of this transit network is facilitated by analyzing 43 topological and operational measures using a coupled map lattice model integrated with a direction-based passenger flow redistribution model. Absorptive capacity thresholds are subsequently identified under different passenger flow-to-route capacity ratios. Finally, the routes are categorized based on their influences on the network robustness using genetic algorithms coupled with $K\text{-}\mathrm{means}$ clustering. The developed approach aims at providing a better understanding of transit systems pre- and postdisruptions by identifying key components that control the network robustness and subsequently devising reliable systemic risk management strategies and recovery plans. Such strategies and plans are expected to facilitate city resilience planning through management of cascade failure risks attributed to natural and anthropogenic hazard events.