Journal of Transportation Engineering, Part A: Systems

There is an increasing interest in signal-timing optimization methods that can consider mobility, safety, and emissions measures simultaneously. The introduction of new models increases the complexity of the required inputs and the relationships between inputs and outputs. This study developed and implemented such a method in an existing computational engine, presenting a sensitivity analysis conducted to provide insight on the effects and order of relevance of 20 key variables on the model’s outcomes and the associated trade-offs among mobility, safety, and emissions. This insight will help the designer, signal control engineer, and traffic analyst when designing intersection geometry and signal control. The statistical analysis of the results showed that the effect of each variable on the overall performance of the model is highly dependent on the combination of other variables. The traffic demand and the size of the intersection, defined by the number of lanes on the arterial, were found to be the most significant variables, affecting all performance measures. Mobility improvement performance usually coincides with emissions improvements, but sometimes occurs at the expense of safety.

This paper investigates the free-flow condition suitable for four-lane divided highways and proposes spatial free-flow speed–based operating speed prediction models. The 85th percentile speed ($V_{85}$) of spatially evaluated free-flow vehicles (passenger car, light commercial vehicle, and heavy commercial vehicle) at the curve center were analyzed with respect to various geometric parameters to develop these models. Among various multiple linear regression techniques explored, the backward elimination method was found to be the best for the adopted data. The curve length ($CL$), deflection angle ($\mathrm{\Delta }$), and preceding tangent length ($PTL$) were identified as the explanatory variables of the developed models. These models were validated for three different sites located in the western and northeastern parts of India. The explanatory variables $\mathrm{\Delta }$ and $CL$ can indicate curve appearance; this is in conformation with the design guideline’s recommendation of adjusting $\mathrm{\Delta }$ and $CL$ for a directional and smooth appearance. Therefore, the proposed models can be used to integrate the quantitative and qualitative guidelines for horizontal curve design. Further, they support the design guidelines’ recommendation to avoid a longer tangent prior to a sharp curve. Sensitivity analysis indicated the passenger car (PC) as the most critical vehicle category. Hence, the speed prediction model of PC was used for developing the nomograms for consistency-based geometric design of horizontal curves along four-lane divided highways.

In this study, changes in traffic flow characteristics were examined in a mixed system consisting of human drivers and autonomous vehicles. In the study, it is accepted that in the environment where the autonomous vehicles are presented, neither vehicle to vehicle communication (V2V) nor the environment and infrastructure communication of vehicles (V2I) exist. The behavioral characteristics of human drivers (reaction time, acceleration, and deceleration, etc.) are derived from data collected from field studies, while the characteristics of autonomous vehicles are determined on the basis of assumptions in literature. Using these data together, the behaviors of the different driver profiles and the movements of the autonomous vehicles under the same conditions are compared by using the SIDRA TRIP program. By using the outcomes of the program and producing different scenenarios, a passenger car equivalence value for autonomous vehicles was tried to be obtained. In the last section of the study, the effect of autonomous vehicles on signalized intersection performance in an unconnected environment is discussed. The results show that autonomous vehicles may significantly decrease intersection capacity.

At the midblock crosswalks with either Pedestrian Hybrid Beacons (PHBs) or Traffic Control Signals (TCSs), pedestrians’ crossing compliance can be considered in terms of the space (crossing location with respect to stripes) and time (waiting for a WALK signal). In any crossing incident, two combinations of a scenario are performed jointly; pedestrians can cross within stripes when WALK signal is either active or inactive, also they can cross outside the stripes when WALK signal is either active or inactive. This study presents the assessment of alternative methodologies for modeling the spatiotemporal crossing compliance of pedestrians. It uses data collected from five signalized crosswalks located along four major arterials in Las Vegas, Nevada. Three models, multinomial logit, ordered logit, and logistic regression (LR), are proposed and evaluated. Bayesian information criterion (BIC), Akaike information criterion (AIC), and misclassification error are the three performance measures used to compare the models. Based on these performance measures, the logistic regression outperformed the other two, as it had low AIC and BIC, as well as low misclassification error. This model was then used to evaluate the factors associated with the pedestrians’ spatiotemporal crossing compliance. The logistic regression results revealed that the active WALK sign and the crossing incidences involve female(s) only are positively associated with pedestrians’ spatiotemporal crossing compliance. On the other hand, the optional one/two cross stages, pedestrian wait time, children and teens, as well as people who cross while riding a bike are negatively associated with spatiotemporal crossing compliance.

This paper investigated the effect of a variable message sign (VMS) on drivers’ route choice behavior through a questionnaire survey and field-observed route choice data. A logistic model was developed based on the questionnaire survey data to reveal the relationship between drivers’ route choice behavior and different attributes as well as traffic information displayed on the VMS. Then, field-collected data were employed to verify the modeling results. It was concluded that when VMS displayed “moderate traffic congestion,” a driver’s average number of daily driving trips and degree of attention to VMS affected their route choice decision. When VMS displayed “heavy traffic congestion,” the impact factors included gender, age, average number of daily driving trips, and degree of driver’s attention to VMS. Modeling results also showed that with the increase of age, the probability of route changing indicated a decreasing trend; drivers who always or occasionally paid attention to VMS preferred to change the original route. In addition, independence test results for field-collected data suggested that the traffic condition information displayed on VMS had significant impacts on drivers’ route choice behavior.

It is a fundamental and important task to determine the minimum fleet size required for a public transit (PT) system because the fleet cost is a major operating expense experienced by PT agencies. More importantly, the fleet size information has been incorporated into various PT operations planning activities. Therefore, effective, efficient, and practical methods as well as tools are required to calculate the minimum fleet size required for a PT system. This work first presents the limitations of a traditional fleet size model that is widely used in the literature. Second, the traditional deficit function (DF)-based fleet size model, which can overcome these limitations, is introduced. However, one restriction of the traditional DF-based fleet size model is that a DF has been proven to be not differentiable at its significant points. To overcome this restriction, this work proposes the idea of using continuous approximation (CA) instead of the step-function feature of the DF model. A numerical example is provided to illustrate the CA of the DFs technique followed by some theoretical results. The effectiveness of the CA of the DFs technique in calculating the fleet size required is demonstrated in a case study of an autonomous modular PT system that is currently being developed in Singapore. Finally, some further extensions and potential applications of the CA of the DFs technique are discussed in connection with a possible design of future urban mobility systems, such as on-demand transit systems, autonomous modular transit systems, and shared autonomous vehicle systems.

Queue length is the most important traffic evaluation index used for traffic signal control at signalized intersections. Most previous studies focused on estimating the queue length in homogeneous traffic flow, but they ignored the difference of the vehicle arrival and the change of traffic flow parameters in heterogeneous traffic flow. In this paper, the authors first use the mixed platoon dispersion model (MPDM) and modify it based on the variation of queue length to estimate the arrival of cars and buses. Furthermore, the heterogeneous traffic flow parameters are analyzed using the heterogeneous traffic flow model and the queue length is estimated with the shock waves of the mixed classes. This approach fully describes the relationship between the disparate upstream traffic arrivals (due to vehicles making different turns) and the variation of the incremental queue accumulation (IQA), and makes up for the shortcomings of the uniform arrival assumption in previous research. The model was tested in a field experiment, the results of which show that the model has satisfactory accuracy and robustness. In addition, its precision is higher than that of the queue length estimation using a single-class model that uses the passenger car unit (PCU) approach to capture the heterogeneity. The limitations of the proposed model are also discussed in this paper.

In current motorway design practice, microscopic simulation is used to assess the traffic safety and capacity implications of design variants. Many different simulation packages are available, and many researchers have invested effort in improving and calibrating models that describe driving behavior. This paper examines traffic simulation models for merging situations in high traffic volume. Based on a multicriteria analysis, two simulation software packages (VISSIM version 10 and MOTUS) were selected and calibrated using a recently collected set of rich empirical trajectory data from an on-ramp, an off-ramp, and two weaving segments in the Netherlands. The results show that both packages are as yet unable to simulate turbulent traffic around motorway ramps realistically, in terms of lane change locations and headway distribution. This is a drawback when assessing traffic operations for specific designs. The simulated gap acceptance distributions were found to have comparable shapes as they were derived from empirical data. This is an advantage when assessing the traffic safety of a specific design.

To realize an early warning system for crowded subway passenger flow, a new calculation method of subway station distribution capacity (SSDC) in real time is proposed. First, a new concept of SSDC is defined according to the gathering and scattering process. Based on automatic fare collection (AFC) data and passenger walking time at key facilities, a real-time calculation model of SSDC has been created. Second, passenger travel characteristics are analyzed by examining two aspects: spatial distribution expressed by cross-section passenger flow and temporal distribution presented by inbound passenger flow each hour. Then, a case study of Zhujianglu Station on Nanjing Metro Line 1 shows the calculation process and verifies the validity of this method. The verification process includes two comparisons: one is to compare the SSDC and station service capacity to obtain possible crowded passenger flow, and the other is to compare the possible crowded passenger flow and the crowded data provided by Nanjing Metro. The same method could be suitable for other metro companies equipped with AFC systems.

Planning and optimizing maintenance actions for rail failures are possible through understanding rail behaviors, which can be captured by statistical models. In this study, a degradation analysis of railroad tracks is presented for three types of geometry failure modes. Using Burlington Northern Santa Fe (BNSF) data set, effective degradation stress factors are identified and inspection intervals are studied to attenuate the effect of hidden maintenance actions. Parameters of a general log linear Weibull model have been determined to estimate the failure probability at different stresses. To check the prediction performance, the models were applied on an independent test set, and the corresponding area under curve (AUC) were reported. Finally, a comprehensive sensitivity analysis was implemented to check the effect of different stress factors on residual life of a defected rail. This type of analysis is crucial in alerting railroad companies to fix a defect and to schedule preventive maintenance actions to reduce the probability of a track failure.

Passengers may change their original travel plans when unplanned metro service disruption occurs, which causes excessive crowding and endangers the safety of passengers. However, it’s not clear about the mechanism of passengers’ behaviors under unplanned service disruption and what factors affect mode shift and travel plan choice behavior. Based on the stated preference data in Guangzhou, China, this paper analyzes the travel plan choice behavior under unplanned service disruption by using a nested logit model. The nested structure consists of two levels: the upper level represents the mode shift choice, and the lower level shows travel plan choice corresponding to the mode shift or not. The results indicate that service disruption attributes and individual attributes are significant predictors to the mode shift. Furthermore, passengers are more sensitive to intermode transfers compared with interline transfers when disruption occurs. The proposed model is also applied to forecast passenger flow volume of metro stations under disruption, and the prediction result indicates that the proposed model well captures passenger mode shift and travel plan choice behavior under unplanned service disruption.

A major hurdle in freight demand modeling has always been the lack of adequate data on freight movements for different industry sectors for planning applications. Both Freight Analysis Framework (FAF) and Transearch (TS) databases contain annualized commodity flow data. However, the representations of commodity flows in the two databases are inherently different. FAF flows represent estimated transportation network flows, whereas TS flows represent production–consumption commodity flows. This study aims to develop a fused database from FAF and TS to determine transportation network flows at a fine spatial resolution (county level) while accommodating production and consumption behavioral trends (provided by TS). To this end, a joint econometric model framework embedded within a network flow approach and grounded in a maximum-likelihood technique is formulated to estimate county-level commodity flows. The algorithm is implemented for the commodity flow information from 2012 FAF data and 2011 TS databases to generate transportation network flows for 67 counties in Florida. The proposed approach may be able to circumvent the need to purchase expensive TS databases in the future.

The popularity and rapid growth of roadside electronic billboards in recent decades has resulted in severe safety issues in China, especially at nighttime. However, in-depth investigation of their impact on driver visual properties has been rather limited. This study intends to bridge this gap and gain insights into the issue based on field experiments conducted on road segments of Freeway G60 and Freeway G2 in China. Fifty-five participants were recruited to drive along the selected freeway segments. Multiple properties of the participants’ visual properties were captured, including eye movement (the transition between regions of interest), glance duration, and the rate of pupil area change. Vehicle behavior parameters, including the standard deviation of lane position and the speed, were also collected in order to estimate the potential risks. Results show that billboards along the freeway result in driver distraction, which is associated with a higher risk to driver and passenger safety. Furthermore, these results indicate that the random and messy billboard installations along freeways in China, especially at nighttime, are a safety concern that needs further exploration.

The congestion caused by primary traffic crashes often exposes subsequent vehicles to the risk of secondary crashes. There has been a proliferation of studies attempting to identify secondary crashes. These studies can be broadly classified into four types: static spatiotemporal threshold–based approaches, queuing model–based approaches, speed contour plot–based approaches, and shockwave-based approaches. However, their respective limitations would lead to the misidentification of secondary crashes. This paper proposes an approach to identifying secondary crashes on freeways by leveraging the spatiotemporal evolution of shockwaves in speed contour plots. In this approach, a spatiotemporal speed matrix is first constructed according to which the speed contour can be plotted. Then, an integer programming model is developed to determine the spatiotemporal region in which a primary crash occurred. The model can ensure that the crash region is consistent with the spatiotemporal evolution of shockwaves. Then, a new way to identify secondary crashes in a speed contour plot is proposed. The model is also extended to handle scenarios in which there are multiple primary crashes in a speed contour plot. Using real data in Beijing, case studies are conducted to validate the proposed approach. The results show that this approach is capable of reducing the misidentification of secondary crashes compared to an approach using static spatiotemporal thresholds and a current identification method based on speed contour plots.

Timetable and vehicle scheduling are important for transit operations. Electric buses are more environmentally friendly compared with conventional buses, and have been developing rapidly and may replace conventional buses in many cities worldwide. This paper focuses on the bus timetabling and vehicle scheduling problem for electric buses and develops a multiobjective optimization model for a single bus line operated with electric buses. The objectives include smoothing the vehicle departure intervals and minimizing the number of vehicles and total charging costs. The constraints reflect the limitations related to the range of departure intervals during different time periods, the vehicle operation mileage, and the charging conditions. A multiobjective particle swarm optimization (MOPSO) algorithm is developed to get the Pareto-optimal solution set. Considering the priority relationships among three objective components, an optimal solution selection strategy is developed. Compared to both the existing schedule and sequential schedule, the integrated model can not only effectively and efficiently reduce the number of vehicles and total charging costs, but also noticeably increase the smoothness of departure intervals. Moreover, it can allow the vehicle charging periods to be evenly distributed during off-peak hours.

When concrete railway crossties are installed in North American freight track and subjected to flexural loads, center negative bending is one of the most critical demands. However, the ultimate flexural capacity at the crosstie center is often unknown and hard to obtain. Because railroads do not always know what the remaining flexural capacity of concrete crossties is, it becomes difficult to assess whether crossties should be removed from service or if it is safe to increase axle loads as an example. To address this challenge, we present a predictive mathematical model based on laboratory experimentation data of various common pretensioned concrete crosstie designs to estimate their center negative bending strength. The model is developed using least absolute shrinkage and selection operator (LASSO) techniques. The final derived equation uses predictor variables that are easily interpreted and applied, and the results are adequate for approximations when limited information is available about the crossties’ characteristics and lengthy structural calculations or additional laboratory testing is not practical. For the investigated crosstie designs, the maximum prediction error was 5.5%.