Journal of Transportation Engineering, Part A: Systems

Connected autonomous vehicles (CAVs) may be able to operate with less longitudinal and lateral spacing than traditional human-driven vehicles (HVs) due to fast and precise control technologies and cooperative maneuvers. With this appealing feature, it is possible to allocate specific narrower highway lanes to CAVs to increase traffic throughput. This paper proposes an analytical lane management framework that determines the optimal number of CAV lanes needed for a highway segment to maximize its throughput considering the narrowed width of CAV lanes. The proposed optimization model investigates three types of vehicles, including CAVs and human-driven light-duty and human-driven heavy-duty vehicles. It takes into account varying mixed-traffic demand levels, CAV market penetration rates, platooning intensities, and CAV technology scenarios. The results from the numerical experiments reveal that the proposed lane management framework with the narrowed CAV lane width increases highway throughput for various parameter settings in different CAV technology scenarios. In order to bring the developed lane management model to the implementation stage, this paper proposes an analytical methodology on how to estimate model parameters (e.g., CAV market penetration rate, platooning intensity, and average headway) with mixed-traffic trajectories when they become available in the near future. We illustrate the application of the developed calibration method with synthetic mixed-traffic data adapted from a data set.

Recently, high-resolution data has emerged as an economical means of obtaining detailed information about signal operations. This has enabled the development of metrics to examine progression, such as cyclic flow profiles similar to those used in the TRANSYT model. Those profiles could further be used to predict changes to traffic flows under trial offset adjustments, enabling offset optimization. An algorithm to do so, called “Link Pivot,” has been used on numerous signalized arterials in Indiana and elsewhere over the past 8 years. This paper reviews the success of the predicted changes through statistical analysis. A linear regression of 569 predicted and actual percent on green values yields an $R^2$ of 0.90. Next, predicted and actual arrival profiles were compared for 357 approaches for which detailed data was available. This analysis finds that the prediction method was successful in about 95% of the approaches considered in this evaluation. The paper dissects the remaining cases and discusses how they could be accommodated.

Video detection system (VDS) sensors play an important role in monitoring traffic flows. Because of limited numbers of camera installations, there is a dilemma in positioning camera sensors between the system’s coverage and its capability to differentiate flows on the network. Existing studies have mainly focused on the observability, evaluation, and estimation of traffic flows in the network sensors location problem (NSLP) field. The effect of covering and differentiating route flows for VDS’s layout has not been fully investigated. This study explored three sensor positioning principles: covering first, differentiating first, and weighted. Corresponding greedy algorithms were developed for each of the three principles. The algorithms using the three principles were tested on a commonly used (by previous studies) medium-sized network and a large-scale real-world network. The covering first principle was found to be more sensitive to the coverage rate of networks. To achieve full coverage, VDS sensors need to be installed on a minimum of 10% and 15.3% of links for medium- and large-scale networks, respectively. The differentiating first and weighted principles outperformed the covering first principle in the large network because they extracted an extra 19% of unique route flows. Therefore, the covering first principle is recommended for deployment of a VDS on small- and medium-sized networks because of its greater capacity to cover route flows and differentiating capacity similar to that of the other principles. It is practical for large-scale networks to utilize the differentiating first and weighted principles to locate VDS sensors because of their greater capacity to extract unique route flows and similar coverage.

Feeder-bus networks between urban rail transit systems and bus stops are inefficient in terms of the cost incurred by passengers, the time taken to reach the destination, and the number and frequency of feeder buses. To solve these issues, a new split delivery model using genetic algorithm (GA) for the feeder-bus network design problem (FBNDP), which is based on the transfer network, is proposed. To accomplish this, the general many-to-one assumption between bus stops and routes is discarded and replaced by the many-to-many assumption, wherein multiple bus stops are served by multiple routes, while considering a new split delivery method. This model is called the feeder-bus network design problem with split delivery (FBNDP-SD). The use of GA and the many-to-many assumption through repeat bus stops resulted in the effective distribution of passenger demand among different routes, reducing the passengers’ generalized travel expenses and optimizing the network structure. While the construction of the FBNDP-SD solution, as well as the interconversion of the FBNDP solution and FBNDP-SD solution, is realized, this paper provides a new research direction of FBNDP.

Emissions reduction has emerged as one of the principal targets in the planning and designing of road alignment today, and intelligent design methods can help optimize road alignment as a response toward more sustainable road infrastructures. The largest share of emissions in road transport occurs in the use phase; hence, considering vehicles’ behavior already in the early stages of the planning process is crucial. This study compares earthwork costs, fuel costs, and tank-to-wheel emissions of alternative road vertical alignments using a spline linear programming (LP) optimization method. The traditional minimal earthwork cost model is tailored and augmented with a fuel item to account for vehicle fuel costs. Three options are considered, including an earthwork-based (EW) optimal road alignment, a balanced earthwork-and-fuel cost (EW-FC) optimal alignment, and a minimal fuel cost (FC) alignment. Calculations are done for a reference test heavy-duty vehicle assumed to operate at uniform speed. The results exhibited that, although leading to some increase in earthwork costs, a design for balanced EW-FC cost yields substantial fuel budget and related emissions savings.

Approximately 47% of fatal crashes in the US and 36% in the European Union occur during the night. Reduced sight distance may contribute to this phenomenon and similar speeds driven between day and night can become problematic along horizontal curves. In this case, drivers are not provided adequate time to adjust their speed for safe negotiation of the curve which can be completed under daylight conditions. This study is the first step toward demonstrating the influence of curvature on nighttime crashes through a preliminary analysis aiming to identify the magnitude of the problem. Data from the US and Greece have been examined and the findings indicate that there is indeed an increased crash occurrence during the night related to the curve radius. Curves with small radii showed an increase in crash occurrence and rate compared to that of corresponding daytime data. The radii of successive curves were utilized as a measure of design consistency and the data again noted differences in crashes between daytime and nighttime conditions.

The paper focused on the issue of curving railroad turnouts in order to use them to connect parallel tracks located in a circular curve on a double-track line. An analytical method of recording was applied, which gave greater application opportunities in particular applications. The basic assumption adopted was to eliminate the reverse curve from the geometrical layout. This indicated the need to adopt a solution with two one-sided curved turnouts. At the same time, in the area where the ends of both turnouts were connected, the values of the occurring bends had to be minimized. After proper curving of the turnouts which were to be applied in both the external and the internal tracks, the question of determining their location becomes the basic issue. The present solution to the problem proposes an appropriate course of action leading to determining the most advantageous mutual position of both turnouts, wherein the connection of the end of the diverging tracks is obtained by means of a straight insert or a segment of a circular arc and a straight insert.

It is necessary to consider both the reservation and real-time travel demand in order to coordinate and optimize vehicle routing and scheduling. In this paper, a two-step coordinated optimization methodology is proposed for a multivehicle responsive transit route and scheduling under mixed demand conditions. At first, a coordinated optimization model of vehicle running route and schedule is constructed to compute the initial route and departure time for each feeder bus service under the reservation demand. The second step builds up a rule-based insertion method to determine whether to respond to the current real-time application according to the passenger’s real-time application. By obtaining the response status of the real-time application, the remaining route for the current bus fleet is determined by using the optimization method, and then the residual is determined through inserting the real-time application. In this study, two experiments will be conducted and tested in order to verify our theories. The first one with fixed and varied departure intervals is verified while simultaneously optimizing the determination of the running route and departure time for each feeder bus under mixed demands. On the other hand, the second one is mainly focused on vehicle ownership for the system, which proves this method can optimize the vehicle inventory to determine the system demand.

Signal-coordinated control systems have been widely implemented on urban arterials. By synchronizing consecutive intersections, signal coordination can significantly improve the throughput of vehicles along arterials. Considerable research has been dedicated to assessing the efficiency of coordinated signalization, while from a safety perspective there is a lack of necessary effort to identify the correlated heterogeneity between injury and property damage only (PDO) crashes. In this paper, the authors adopt a multivariate Poisson log-normal (MPLN) model to reveal the heterogeneous connectivity along coordinated arterials. In addition, spatial correlations of conditional autoregression (CAR) and multivariate conditional autoregression (MCAR) are added to the MPLN model for the purpose of calibrating the regressive results. With the surveyed arterials in Ann Arbor, Michigan, the case study shows that (1) the MPLN-MCAR model predicts the results more accurately compared to the MPLN-CAR and MPLN models, (2) the spatial correlation of injury crashes differs significantly from that of PDO crashes, and (3) there is a strongly correlated heterogeneity between injury and PDO crashes. The findings serve to provide effective countermeasures for safety planning, design, and management of signal-coordinated arterials.

Traditional travel modes have been slowly losing their market share to e-hailing services, especially among younger generations. This paper examines characteristics of young travelers’ travel mode choice, with various e-hailing modes described and modeled. A stated preference survey was conducted in Nanjing, China. A total of 314 valid responses were collected from young travelers from 16 to 35 years old. Descriptive statistics illustrate the transition details of the mode choices for interviewees of different genders, income levels, and educational levels. A cross-nested logit (CNL) model was employed to analyze the relationship between impact factor and young travelers’ mode choice behaviors, compared with two nested logit models. Factors, including personal profile and different scenario conditions, were examined. Comparison results indicated that the CNL model was statistically superior to the other two models. It was found that young travelers who were male, higher education level, lower income, rigid trip purpose, and long distance were more likely to choose e-hailing services. The modeling results also revealed the expected negative impacts to factors such as male, lower education, rigid trip purpose, and long distance on bus service, and adverse weather on taxi service.

Traffic crashes occur frequently in expressway diverging areas. The influential factors for traffic crashes in expressway diverging areas were investigated in this study using vehicle trajectory data. An hourly conflict risk index (HCRI) was introduced to develop a traffic conflict prediction model for expressway diverging areas. The data collectors were trained to identify conflict severity, and Tracker 5.0 was used to calculate the time to collision (TTC) for rear-end and lane-change conflicts, respectively. Based on the value of direct economic losses, the traffic risk index for traffic conflicts of different types and severities is established, and the severity of traffic conflict was characterized by the HCRI. A multivariate linear regression model was applied to analyze the relationship between HCRI and various influential factors. A comparison between an hourly conflict ratio (HCR) model and the HCRI model showed that the HCRI model performed better. Finally, it was found that the mainline traffic volume and the ramp traffic volume were positively associated with HCRI, while the deceleration lane length and mainline speed were negatively associated with the HCRI.

This study applies the full velocity difference (FVD) model to investigate commuters’ driving process in a speed limit corridor with a toll station and explains how some factors affect the outflow of the traffic corridor system from the microscopic perspective. First, a micromodel system is constructed to depict the commute process in a traffic corridor with a toll station considering the speed limit. Second, a metastable state is defined in the paper for deriving the properties of commute problem conveniently. Finally, numerical tests are conducted to investigate some factors’ effects on traffic properties. This study concludes that many traffic properties will be convergent when the commuter’s number increases, and that the length of the road, the speed limit, and the service time of the toll station have effects on the time headway of the origin at the equilibrium state. The results show how the outflow forms in a speed-limit corridor with a toll station from the microscopic perspective and provide insights for policymaking.

Twist is a track defect that affects comfort by inducing jerk and rolling and endangers the derailment safety of a vehicle by wheel off-loading. Currently, there is no formula in the literature that relates speed and jerk with twist. In this paper, a relationship between jerk, speed, and twist is developed. The relationship was used to assess the tolerance values of twist for design, construction, and maintenance of the track. The relation was also used to assess slow speed on a twisted track beyond the maintenance limit. The advantage of the formulas is that they are speed specific. The relation between twist, jerk, and speed was theoretically validated. The derived equations from the aforementioned relation to estimate tolerance values of twist were validated by comparing with the values given by US and European specifications. A literature review and analysis was performed on chord length to measure the twist. A chord equal to the wheel base is suggested to measure twist on a newly constructed or renewed track. On a revenue track, chord lengths equal to both wheel base and truck center to center distance are recommended to capture both short- and long-wave twist defects. The suggested chord lengths would integrate vehicle parameters, track defects, comfort, and safety.

Passenger ride comfort (PRC) is one of the most important performance indicators of railway transportation. The current methods for computation and evaluation of railway ride comfort require measurement of train accelerations and dynamic vehicle–track interaction parameters, which is costly and sometimes impractical. Despite the importance of passenger ride comfort in design and operation of railways, there is a lack a reliable PRC prediction model in the available literature. Addressing this limitation, an effective and practical PRC prediction model/index is established in this study, taking into consideration track and rolling stock influencing parameters for the first time. For this purpose, correlations were developed between PRC levels and track geometry parameters as well as rolling stock dynamic properties. The PRC level was computed based on accelerations obtained from accelerometers installed on the wagon floor. The track geometry parameters were obtained from a track recording car. Practicability and reliability of the prediction model were discussed by applying the model in a railway line. A good agreement was shown between the PRC levels obtained from the prediction model and those of field measurements. Application of the prediction model in the real world of railway engineering is illustrated.