Search

Public transport in rural areas is under pressure because demand is low and dispersed. To reduce costs, flexible and on-demand services are often proposed as alternatives for conventional bus services. Conventional services are generally not suitable for rural areas, because the demand is low and dispersed. In this paper, a stated preference survey is designed to identify the preferences of rural bus users for alternative services. Other than the traditional bus, two other modes are included in this study: a demand responsive transport (DRT) service and an express bus service with bike-sharing services for last mile transport. Given the on-demand nature of these alternatives, flexibility- and reliability-related attributes are included in the stated preference survey. The results from the choice model indicate that the reliability and flexibility aspects do not have a large effect on the preference for the on-demand alternatives. Instead, cost, access and egress times, and in-vehicle time play a bigger role in individuals’ preferences toward the different alternatives. A sensitivity analysis shows that changes in the operational characteristics can make the on-demand alternatives more attractive. However, many bus users still prefer the conventional bus service over the on-demand alternatives.

Traffic engineers rely on microscopic traffic models to design, plan, and operate a wide range of traffic applications. Recently, large data sets, yet incomplete and from small space regions, are becoming available thanks to technology improvements and governmental efforts. With this study we aim to gain new empirical insights into longitudinal driving behavior and to formulate a model which can benefit from these new challenging data sources. This paper proposes an application of an existing formulation, Gaussian process regression (GPR), to describe individual longitudinal driving behavior of drivers. The method integrates a parametric and a non-parametric mathematical formulation. The model predicts individual driver’s acceleration given a set of variables. It uses the GPR to make predictions when there exists correlation between new input and the training data set. The data-driven model benefits from a large training data set to capture all driver longitudinal behavior, which would be difficult to fit in fixed parametric equation(s). The methodology allows us to train models with new variables without the need of altering the model formulation. And importantly, the model also uses existing traditional parametric car-following models to predict acceleration when no similar situations are found in the training data set. A case study using radar data in an urban environment shows that a hybrid model performs better than parametric model alone and suggests that traffic light status over time influences drivers’ acceleration. This methodology can help engineers to use large data sets and to find new variables to describe traffic behavior.

Detector data can be used to construct cumulative flow curves, which in turn can be used to estimate the traffic state. However, this approach is subject to the cumulative error problem. Multiple studies propose to mitigate the cumulative error problem using probe trajectory data. These studies often assume “no overtaking” and thus that the cumulative flow is zero over probe trajectories. However, in multi-lane traffic this assumption is often violated. Therefore, we present an approach to estimate the change in cumulative flow along probe trajectories between detectors based on disaggregated detector data. The approach is tested with empirical data and in microsimulation. This shows that the approach is a clear improvement over assuming “no overtaking” in free-flow conditions. However, the benefits are not clear in varying traffic conditions. The approach can be applied in practice to mitigate the cumulative error problem and estimate the traffic state based on the resulting cumulative flow curves. As the performance of the approach depends on the changes in traffic conditions, it is suggested to use the probe speed observations between detectors to assign an uncertainty to the change in cumulative flow estimates. Furthermore, a potential option for future work is to use more elaborate schemes to estimate the probe relative flow between detectors, which may, for instance, combine probe speeds with estimates of the macroscopic states along the probe trajectory. If these macroscopic estimates are based on the cumulative flow curves at the detector locations, this would result in an iterative approach.

Cooperative (automated) vehicles have the potential to enhance traffic efficiency and fuel economy on urban roads, especially at signalized intersections. An optimal control approach to optimize the trajectories of cooperative vehicles at fixed-timing signalized intersections along an arterial is presented. The proposed approach aims to optimize throughput first, and then to maximize comfort while minimizing travel delay and fuel consumption. The proposed approach is flexible in dealing with both quadratic and more complex cost functions. Assuming fixed timing signal control in a cycle and vehicle-to-infrastructure communication, the red phase is taken into account in position constraints for vehicles that cannot pass the intersection in the green phase. Safety is guaranteed by constraining the inter-vehicle distance larger than some desired value. The approach is scalable and can be used for joint trajectory planning of one platoon approaching another stationary platoon. It can also be extended to multiple intersections with fixed signal plans. To verify the performances of the controlled platoon, simulation under three different traffic scenarios is conducted, namely: an isolated intersection with/without downstream vehicle queues, and platoon control at multiple intersections. Three baseline scenarios without control are also designed to compare performances in relation to both mobility and fuel consumption in each controlled scenario. The results demonstrate that the controlled vehicles generate plausible behavior under control objectives and constraints. Moreover, the consideration of downstream vehicle queues and the application at both an isolated signalized intersection and arterial corridors on urban roads verify the flexible characteristics of the control framework.

Demand-responsive transport (DRT) services (collective on-demand services, such as shared ridesourcing and microtransit) offer a collective flexible travel alternative that can potentially complement fixed transit (FT). The combination of an on-demand and line-based service holds the promise of improved mobility and increased service coverage. However, to date, it remains unknown whether DRT services deliver these much anticipated improvements. This study presents an assessment framework to evaluate the performance of DRT and related changes in accessibility, and performs an empirical analysis for a recently introduced DRT service in the Netherlands. The framework includes a performance benchmark between DRT and FT based on the computation of generalized journey times of the DRT rides and the FT alternatives, and can help identify whether DRT is used as a complement or a substitute for FT. The framework covers the spatial and temporal dimensions, and the explicit consideration of rejected trips is an integral part of the evaluation. Results suggest large accessibility improvements for DRT users, especially for some underserved origin–destination pairs.

In this study the influence of interaction characteristics on operational walking dynamics within a crowd—specifically the influence of the distance headway, time headway, angle of sight, angle of interaction, walking speed, and number of pedestrians located nearby on a pedestrian’s change in speed and direction—is investigated. To the authors’ best knowledge, this is the first time that the combined effect of the characteristics of interactions between pedestrians on the operational walking dynamics of pedestrians has been quantified. The walking speed and the number of pedestrians in close proximity were found to influence the adaptation of speed and direction. The other characteristics of the interaction affect either the change in speed (i.e., distance headway and interaction angle) or the change in direction (i.e., time headway and angle of sight). The results of this study strongly indicate that the density experienced by pedestrians is not the only characteristic of the crowd that affects pedestrians’ operational walking dynamics. Consequently, to model crowd movements correctly, the other characteristics of the interaction must also be taken into account in pedestrian flow theory and simulation models.

Ramp metering is an effective means to reduce freeway bottleneck delay that results from the capacity drop phenomenon when congestion sets in. The bottleneck flow is kept at free-flow capacity by temporarily storing vehicles traveling toward the bottleneck at the ramp. The metering duration is normally limited because of the finite amount of ramp storage space available to prevent undesired spill back to the urban network. A beneficial extension of the metering duration might be achieved by strategically choosing upstream intersection arms that reduce their inflow to the ramp. For that purpose, coordination needs to be realized between the ramp meter and its intersection controllers located upstream, which will also hinder vehicles not traveling toward the ramp. In this contribution, an evaluation approach is put forward to decide objectively which intersection buffers (arms) should be included in the coordination. To quantify the resulting delays in the system, cumulative inflow and outflow curves are developed as a function of the involved situation-specific variables. This approach enables one to determine the optimal set of coordinated buffers beforehand and to gain insight into the effect of the various system variables on the delays. By means of worked examples, these effects are illustrated, and the way to determine the optimal set of coordinated buffers is shown. Results show that the length of the peak period and the size of the capacity drop strongly determine the coordination benefits, and, hence, a buffer’s minimum required fraction of traffic traveling past the bottleneck needed to result in beneficial coordination.

In this paper, a coordinated ramp-metering approach is proposed; the approach is based on the synchronization of the time it takes for an on-ramp to run out of space. The time that congestion and the associated capacity drop can be prevented at a freeway bottleneck by means of ramp metering depends on the available ramp space for temporarily storing the vehicles. For a single on-ramp, the available space and thus the metering time at the bottleneck are often limited. By means of coordination, upstream storage space can be used to prevent a breakdown for a longer period of time. To accomplish that aim, upstream ramps need to reduce their inflow into the mainstream with respect to their own metering task. The extent to which upstream inflow reductions are effectively used to extend the metering time at the bottleneck is determined by the order and timing at which the coordinated ramps run out of space. The proposed approach aims at always saturating the more upstream-located ramps before more downstream-located ones; that is, the ramps run out of space in the downstream direction to make sure that all realized assistance is used by the ramp that is metering at the bottleneck. By means of a simulation test case, the functioning of the algorithm is demonstrated and compared with a no-control scenario, local ramp metering, and an alternative coordination algorithm that aims at filling the ramps equally. Both coordination approaches strongly improve the network performance with respect to local ramp metering, but the proposed coordination approach is able to postpone the breakdown for a longer period of time.

This article reviews the multiscale modeling concept as an adaptive strategy in traffic flow modeling. The central objective for the multi-scale model is to describe and to predict traffic phenomena (e.g., individual accelerations, queuing) that occur at different scales by switching between modeling paradigms. The emphasis in this paper is placed on resolving inconsistencies that arise in the process of integrating models, which are addressed with the mesoscopic Lighthill–Whitham and Richards models—together, widely known as the Lighthill–Whitham–Richards (LWR) model—as the coarse-scale model in the framework. The mesoscopic model is fully consistent at the macroscopic scale with the classic LWR model but does facilitate tracking of individual vehicles. Because of this characteristic, this model can be combined with microscopic models, because no disaggregation or aggregation process is needed. It is thus implied, for example, that travel information (e.g., destination) can be carried over this model without loss of information. Also, vehicle heterogeneity can be incorporated in the mesoscopic model and, as a result, is preserved in the process of switching between models. A brief overview of the mesoscopic LWR model and its solution by the variational theory is presented. General boundary conditions, including internal boundaries (e.g., moving bottlenecks) and interfaces between models, are analyzed and evaluated. The solution of some cases, in which the stationary and moving bottlenecks interact, is provided. The paper concludes with a case study that demonstrates the integration of microscopic and mesoscopic models in a multiscale model and investigates the influence of various boundary conditions on traffic features.

Automated vehicles are expected to have a substantial impact on traffic flow efficiency, safety levels, and levels of emissions. However, field operational tests suggest that drivers may prefer to disengage adaptive cruise control (ACC) and resume manual control in dense traffic conditions and for maneuvers such as changing lanes. These so-called authority transitions can have substantial effects on traffic flow. To gain insight into these effects, a better understanding is needed of the relationships between these transitions, longitudinal dynamics of vehicles, and behavioral adaptations of drivers. In this context, a driving simulator experiment was set up to gain insight into the effects of authority transitions between ACC and manual driving on longitudinal dynamics of vehicles. Participants were assigned randomly to one of three conditions. In the control condition, participants drove manually. In the first experimental condition, a sensor failure was simulated at a specific location where drivers were expected to resume manual control. In the second experimental condition, drivers switched ACC off and on by pressing a button whenever they desired. Statistical tests indicated that the distributions of speed, acceleration, and time headway differed significantly between the three conditions. In the first experimental condition, the speed dropped after the sensor failure, and the time headway increased after the discretionary reactivation of ACC. These results seem to be consistent with previous findings and suggest that authority transitions between ACC and manual driving may significantly influence the longitudinal dynamics of vehicles and potentially mitigate the expected benefits of ACC on traffic flow efficiency.

A new maritime traffic model describes vessel traffic in ports and inland waterways better. In this research, vessel behavior is categorized into a tactical level (route choice) and an operational level (dynamics of vessel behavior). This new maritime traffic model comprises two parts. The route choice model resulting in the vessel's preferred route and the operational model describing the maneuvering behavior, including interactions between vessels. This paper presents the vessel route choice model, which is based on disutility or cost minimization. The cost is determined by characteristics of the infrastructure, such as expected sailing time and distance to the bank. It is assumed that the bridge team will try to follow a preferred route that minimizes the cost to the destination. To calculate this preferred route, the so-called “value function” is defined as the minimum disutility function in continuous time and space. Subsequently, the value function is solved with dynamic programming and a numerical solution approach. Data of unhindered vessel behavior in the Port of Rotterdam, Netherlands, collected with an automatic identification system, are used to calibrate the vessel route choice model. The calibrated results of the route choice model show plausible preferred routes in the research area, which aid understanding of the desired vessel behavior (route). These results could be used to improve vessel traffic management and provide a basis for predicting vessel behavior at the operational level.

It has been empirically observed for years that the queue discharge rate is lower than the prequeue capacity. This difference is called the capacity drop. The magnitude of capacity drop varies over a wide range, depending on the local traffic conditions. However, it is unknown what determines the capacity drop value. No thorough empirical analysis has yet revealed a reliable relationship between the capacity drop and the congestion level. This paper fills the gap by revealing, through empirical analysis, the relationship between vehicle speed in congestion and the queue discharge rate. The research studies congested states in which speed ranges from 6 to 60 km/h. The queue discharge rate is shown to increase considerably with increasing speed in the congestion. In contrast to previous research, this study bases the relationship on empirical data collected on freeways, and the data present a sufficiently large observation sample. A discussion about the influence of weather and study site characteristics on the discharge rate indicates that the relationship needs site-specific calibrations. This study provides a better prediction of capacity drop and a better theoretical understanding of the fluctuations in capacity drop.

Praktijkproef Amsterdam (Field Operational Test Integrated Network Management Amsterdam) is a project focused on the design and implementation of an innovative system for the coordinated deployment of traffic management measures in the regional network around the city of Amsterdam, Netherlands. On the basis of the generic functional architecture described in previous publications, monitoring and control functions were specified, implemented, tested, and tuned, both in a model environment and in practice. The focus here is on the controller design, which adapts and generalizes the master–slave structure used in the well-known HERO algorithm. It is shown how the principles are applied to use the urban arterials as storage locations for freeway metered traffic as well as to solve problems occurring on the urban arterials themselves—for example, in the case of spillback or gridlock. The key control principles are explained and the resulting feedback controller is analyzed to show how to choose its parameters to ensure stability and minimize the time needed to achieve the target value. To this end, a new methodology is proposed to analyze the dynamics of the controller in relation to the control parameters for different controller designs. Finally, the implementation of the system is discussed, and preliminary tuning and evaluation results from the field tests are provided.

The macroscopic fundamental diagram or network fundamental diagram (NFD) describes the relationship between network flow and the number of travelers within a network. The shape of the NFD is important for control purposes such as perimeter control or routing. However, the influence of network structure on NFD shape is unknown. This topic is studied, along with the influence of adding urban arterials (i.e., decreasing homogeneity but adding road capacity) and the effect of arterial location. For this purpose, a tool is developed to create realistic networks with preset characteristics (i.e., road length and road type). The exact network is created at random. Simulation results indicate that networks with similar characteristics have different NFDs. The tool creates networks that have arterials next to urban roads. Simulation results indicate that for these more heterogeneous networks, spread is higher and maximum production is lower in the NFD. The exact location of a major road crossing the network does n...

Multiscale traffic flow modeling, which combines traffic models of different scales (e.g., of network and traffic representation), offers many opportunities for the simulation and evaluation of traffic measures and policies in large networks. For example, macroscopic continuum models of traffic flow could be used to simulate the propagation of traffic on large freeway corridors and networks efficiently and parsimoniously, whereas microscopic models of traffic flow could be used to simulate those parts of the network where more detail (e.g., lateral behavior) is required, such as at intersections and toll plazas. One key requirement of such a multiscale framework is related to the interface between models of different scales. This interface should ensure consistency in (a) vehicle conservation and (b) vehicle and traveler properties that must be transported over this interface (e.g., vehicle type, occupancy, and destination). In the approach presented, both types of consistency problems can be solved simul...

Automation may be assumed to have a beneficial impact on traffic flow efficiency. However, the relationship between automation and traffic flow efficiency is complex because behavior of road users influences this efficiency as well. This paper reviews what is known about the influence of automation on traffic flow efficiency and behavior of road users, formulates a theoretical framework, and identifies future research needs. It is concluded that automation can be assumed to have an influence on traffic flow efficiency and on the behavior of road users. The research has shortcomings, and in this context directions are formulated for future scientific research on automation in relation to traffic flow efficiency and human behavior.

This paper presents an innovative way to model the decision-making process of the bridge team of a ship. The model aims to provide methods to include human decision making in comprehensive simulation models that can describe the movement of vessels, including hydrodynamic effects; external effects due to wind, current, and waves; waterway geometry; and the interaction with other vessels. The paper uses a simple model to describe a vessel's dynamics and the impact of the control decisions on these dynamics, although generalization to more comprehensive maneuver models is straightforward. The mathematical modeling framework is presented on the basis of a set of behavioral assumptions. The model is described as a differential game in which the bridge team is assumed to react on the expected behavior of other vessels. Different behavioral strategies (risk prone, average risk, and neutral risk) lead to the different models described in the paper. The dynamics of the model are illustrated by simple examples. The results are plausible and clearly show the potential of the approach. The paper offers some direction for future development.

Emergency situations (e.g., evacuations following a disaster) have been shown to affect traffic flow operations substantially. However, the best way to model the adaptation effects in longitudinal driving behavior underlying this impact had not been made clear. Furthermore, the macroscopic consequences of the adaptation effects in longitudinal driving behavior had also not been made clear. This study sought to clarify these modeling issues and macroscopic consequences by estimating parameter values and model performance of the intelligent driver model with the data obtained through a driving simulator study. In addition, this paper presents the results of a case study that used a microscopic simulation program and the parameter values obtained through the estimation of the intelligent driver model. Results show that emergency situations have a substantial influence on parameter values and performance of the intelligent driver model. Furthermore, results show that the adaptation effects represented in parameter values and model performance have a substantial influence on macroscopic flow characteristics. A discussion of results and recommendations for future research are provided.

On Freeway A20 near Rotterdam, Netherlands, a trial with dynamic speed limits began on June 28, 2011. On a 4.2-km stretch, the speed limit increased from 80 to 100 km/h as soon as congestion appeared to set in and at night. The aim of dynamic speed limits was to improve traffic operations and to avoid deterioration in the local air quality. This paper presents an assessment of this trial with respect to traffic operations, air quality, noise level, and traffic safety. Traffic operations on A20 appeared to have improved significantly as a consequence of the dynamic speed limits, which produced a reduction in the number of lost vehicle hours by 600 (20%). This improvement was the result of a 4% increase in the free-flow capacity at the main bottleneck on the freeway stretch. The dynamic speed limits caused a change in driver behavior: the median lane was better occupied when flow increased. Air quality deteriorated slightly. The effects varied along the stretch with a maximum increase in nitrogen oxides and particulate matter10 emissions of 3.7% and 3.6%, respectively. However, the effects on the average concentration of nitrogen oxides per year were limited. The noise level appeared to increase slightly with 0.2 dB. This increase occurred mainly during the two peak periods. The indicators for traffic safety showed sometimes a (possibly) positive and sometimes a (possibly) negative effect. However, it was not likely that dynamic speed limits had a significant negative effect on traffic safety.

Because of an increasing exchange of data between measurement sites, the area over which traffic control is applied is also increasing. This situation leads to three new challenges: (a) working with the large quantities of data (transmit, store), (b) estimating the traffic state, and (c) controlling a large area with many controllers (and thus large solution space). This paper introduces a new way of describing the traffic state for a large area, one that requires much less data and nevertheless gives an accurate representation of the state. The macroscopic fundamental diagram (MFD) links the production (the average flow) to the accumulation (the average number of vehicles) in an area. This paper shows that MFD can be converted to a generalized MFD (GMFD) for urban freeways; the GMFD relates the production to the accumulation and the spatial spread of density. Analysis of 10 months of data from the Amsterdam, Netherlands, ring road freeway showed that GMFD is a continuous function that increases and decreases with accumulation, as does a fundamental diagram, and decreases with the spatial spread of density. The predictive performance of GMFD was tested with a nonparameterized fit and by fitting a functional form; each test performed equally well. Predicting the production is important, especially near the maximum production. GMFD explains much more of the spread in the production than MFD does, especially near this maximum production. Thus, this lean traffic state description can be used in setting a target for traffic control.

This paper develops a simple, robust framework for the problem of finding the route with the least expected travel time from any node to any given destination in a stochastic and time-dependent network. Spatial and temporal link travel time correlations are both considered in the proposed solution, which is based on a dynamic programming approach. In particular, the spatial correlation is represented by a Markovian property of the link states, in which each link is assumed to experience congested or uncongested conditions. The temporal correlation is manifested through the time-dependent expected link travel time given the condition of the link traversed. The framework enables the use of a route guidance system, in which at any decision node within a network, a decision can be made on the basis of current traffic information about which node to take next to achieve the shortest expected travel time to the destination. Numerical examples are presented to illustrate the computational steps involved in the framework to make route choices and to demonstrate the effectiveness of the proposed solution.

Because vessel traffic in ports and waterways is growing quickly, much attention has been given to maritime traffic safety and port capacity. Many simulation models have been used for predicting traffic safety and port capacity in ports and waterways. However, maritime traffic models have considered only a few aspects; the influence on safety of human behavior and external factors has not been included. An analysis based on data from an automatic identification system was performed under various external conditions in an investigation of vessel behavior and external influencing factors. The study area included a junction and a slight bend with high maritime traffic density within the port of Rotterdam, Netherlands. Vessels were classified according to type and gross tonnage. Equidistant cross sections approximately perpendicular to the navigation direction were used for investigation of vessel behavior, including speed, course, and path for each vessel category. The influence of external factors (wind and visibility) on vessel behavior was identified through a comparison with the behavior of unhindered vessels. In the analysis, specific thresholds were set for selecting external conditions and eliminating the influence of encounters. The analysis of unhindered vessels for each vessel category provided insight into vessel behavior. The results revealed that wind had an influence on vessel speed and that visibility affected vessel speed, course, and path. Analysis results can be used as input for the development of a new maritime traffic model, as well as for its verification and validation.

Simulation software packages have been developed to simulate traffic flow operations in emergency situations. However, it is not yet clear which adaptation effects in empirical longitudinal driving behavior can be observed. Furthermore, it is not yet clear to what extent these adaptation effects in emergency situations are represented in current car-following models. A driving simulator experiment was performed to induce longitudinal driving behavior under emergency conditions. Longitudinal driving behavior was measured through registered behavior in the driving simulator and analyzed with a multivariate analysis of variance. A new data analysis technique used so-called action points in psychospacing models. Acceleration and jumps in acceleration at action points were determined with multivariate regression analysis as well. The results show that substantial and significant adaptation effects in longitudinal driving behavior can be observed in an emergency situation. The results also show substantial differences in the position of action points in the speed–spacing plane in the psychospacing model between normal driving conditions and emergency conditions, as well as substantial differences in acceleration and jumps in acceleration at these action points.

Prediction of the time evolution of origin–destination (O-D) matrices is important for many applications in the traffic domain. These applications range from ex ante evaluation to real-time prediction and control. Since O-D matrices are high-dimensional multivariate data structures, both the specification and the estimation of O-D prediction models are methodologically and computationally cumbersome. This paper demonstrates that a significant reduction of the dimensionality of the O-D data that preserves structural patterns can dramatically reduce computational costs without a significant loss of accuracy. This paper explores the application of principal component analysis (PCA) for this purpose. PCA shows that the dimensionality of the time series of O-D demand can be reduced significantly. This paper also shows how the results from the PCA method can be used to reveal the structure in the underlying temporal variability patterns in dynamic O-D matrices. The results indicate three main patterns that can be distinguished in dynamic O-D matrices: structural, structural deviation, and stochastic trend patterns. Insight into how these trends contribute to each O-D pair and how this information can be further used to predict dynamic O-D matrices on the basis of a set of dynamic O-D matrices obtained from real data is provided.

Lane utilization on the highway is affected subtly by dynamic traffic management systems such as speed controls and lane management. To optimize the operation of dynamic traffic management, a better understanding of lane utilization is required, in particular, of how the flows of different vehicle classes (e.g., passenger cars, lorries) vary across the carriageway. Most loop detector systems do not collect this multilane, multiclass count data. This study developed a procedure for estimating multilane, multiclass counts from a variety of standard aggregate loop data formats from around the world. The estimation procedure involved the inference of multilinear regression laws that relate multilane, multiclass data to standard aggregate formats. The regression laws were then trained with small samples of individual vehicle data on a site-by-site basis. Preliminary results showed that the estimation procedure worked rather well, even when the input data were minimal—the extreme case being that of (U.S.-style) single-loop data, for which only flow and occupancy were available on a by lane basis. An error analysis indicated that small amounts of individual vehicle data were sufficient to train the estimator, provided they contained a representative mix of the flow behaviors at the site in question. Further work is required for the practical development of the tool, but it appears to have a wide range of potential uses for both researchers and practitioners.

Emergency doors may be bottlenecks in the evacuation of a building. When a pedestrian simulation model is used to assess designs of buildings, the model should accurately predict the behavior of pedestrians around emergency doors and thus the doors’ capacity. Data from laboratory experiments were used to calibrate the pedestrian simulation model Nomad. In these experiments, large heterogeneous groups of people passed through a door under evacuation conditions. The collected trajectory data were used as input for an automated calibration procedure, which yielded parameter estimates for individual pedestrians. This automated calibration procedure has been extended through inclusion of data from multiple pedestrians into a single estimate. This change overcomes convergence problems because the log likelihood is insensitive to changes in some of the parameters and because of the problems of unrealistic parameter estimates. The resulting parameter distributions provided insight into pedestrian behavior. Dedicated parameter sets for the elderly, adults, and children were estimated. The parameter sets and thus behaviors of the types of pedestrians were compared and showed that different behavior was indeed observed.

Lane changes are an important aspect of freeway flow. Most models of lane change are microscopic. Lane change behavior of individual vehicles or drivers is described, and, therefore, models are calibrated microscopically. Macroscopic validation often is restricted to the distribution of vehicles across lanes. To the best of the authors' knowledge, no systematic analysis has been made of the number of lane changes as a function of the operational characteristics of the origin and target lane. This paper fills the gap in lane change literature with an analysis of the number of lane changes as a function of several incentives. On the basis of data availability, two “simple” sites were selected, that is, as close as possible to a straight continuous freeway. Statistical analysis at the selected sites revealed that drivers changed lanes on average once per 2 km driven. Furthermore, an analysis of the number of lane changes (per kilometer per hour) as a function of the density in the origin lane and in the target lane showed that the number of lane changes increased with the density in the origin lane for a fixed density in the target lane. The number of lane changes also increased with the density in the target lane for a fixed density in the origin lane. The underlying mechanism was therefore different from gap-acceptance theory. The analyses presented in this paper can be used to verify qualitatively (microscopic and macroscopic) lane change models and to propose better ones.

With the reliability of travel time high on the political agenda, tools are needed to predict these indicators for the reliability of travel time in ex ante evaluations. In this paper such a framework is developed for assessing traffic measures and policies under a range of scenarios, with a particular focus on the resulting travel time distribution. The framework uses Monte Carlo sampling to generate stochastic realizations of demand and supply characteristics. The former characteristics relate to variations in travel patterns, the latter to variations in drive behavior (capacities, speeds). An extensive effort is made to account for variability in demand and supply resulting from weather, road work, special events, and so forth. The overall conclusion is that incorporating the variability of demand and supply characteristics has a large effect on the results of evaluation studies. This effect is demonstrated in an example case on a real Dutch traffic network, in which a typical dynamic measure, the opening of a peak hour lane, is evaluated with the framework. In this case, the average demand and supply conditions used for the one-shot evaluation lead to an underestimation of spillback effects, which are captured when evaluating over a wide range of demand and supply characteristics. As a result, the gains in average travel time are significantly underestimated in the one-shot procedure (6% instead of 26%). Further research should focus on improving speed and validity over a wider range of circumstances of the developed framework.

Few Active Traffic Management measures proposed in the past consider the distinction of different vehicle classes. Examples of vehicle-class specific measures are truck lanes and high-occupancy/toll (HOT) lanes. We propose that the distinction of different vehicle classes, with different flow characteristics and societal and economic function, can contribute to the effectiveness of traffic management measures. In this paper, we develop a multi-class controller that reroutes the traffic class-specifically dependent on the vehicle class. The vehicle-class specific properties such as vehicle length and value of time are used in a model-predictive control approach to optimize the total time spent and the economic costs. Experiments in a simple network with synthetic data show that a multi-class controller outperforms a single-class controller. We further show that the value of time and the incident strength have an influence on which vehicle class is rerouted by the multi-class controller.

An excess number of vehicles in a traffic network will reduce traffic performance. This reduction can be avoided by traffic management. In particular, traffic can be routed such that the bottlenecks are not oversaturated. The macroscopic fundamental diagram provides the relation between the number of vehicles and the network performance. One can apply traffic control on this level, in order to overcome computational complexity of network-wide control using traditional control levels of links or vehicles. Main questions in the paper are: (1) how effective is traffic control using aggregate variables compared to using full information and (2) does the shape of the macroscopic fundamental diagram change under traffic control. A grid network with periodic boundary conditions is used as example, and is split up into several subnetworks. The following routing strategies are compared: (1) the shortest path in distance, (2) the path shortest in time (dynamic due to congestion), (3) an approximation of the path shortest in time, but calculated using only variables aggregated for over a subnetwork, (4) an approximation of the path shortest in time, but calculated using only subnetwork accumulation. For routing strategy 3 and 4 only information aggregated over the subnetwork is used. The results show improved traffic flow using detailed information. Effective control is also possible using aggregated information, but only with the right choice of a subnetwork macroscopic fundamental diagram. Furthermore, when optimizing with detailed information – an hence in a subnetwork – the macroscopic fundamental diagram changes.

The realization of traffic management on a network level is not only theoretically complex, but also practically challenging because the traffic management policy of the road authorities must be taken into account. In the Netherlands, this policy harmonizes the interests of involved stakeholders by means of a common vision on the functioning of the network, expressed in road priorities and the corresponding target service levels. As a result, network states that reflect the policy's objectives in a systematic and comprehensible way must be realized. This paper presents a predictive route guidance approach that is able to operationalize the formulated policy. This approach degrades and restores target service levels of routes according to their difference in priority, with respect to the network performance. The control approach consists of a finite state machine that determines target service levels according to predicted traffic conditions. These target service levels are used as setpoints in a feedback controller, and the result is a corresponding output signal of a variable message sign. By means of a test case, the finite state machine is compared with a model predictive route guidance controller (that realizes system optimal conditions) and with a user equilibrium feedback controller (that realizes user optimal conditions). Results showed that the finite state machine was able to prevent or limit the effects of phenomena that caused decreased network performance in a comprehensible and efficient way while also accounting for the interests of the road users.

Adverse weather conditions have a substantial effect on traffic flow. However, the adaptation effects in longitudinal driving behavior that underlie this impact are unclear, as are the determinants. A driving simulator experiment was performed with a repeated-measures design and 25 participants. The adaptation effects in actual longitudinal driving behavior and the physiological indicators of mental workload (i.e., heart rate and heart rate variability) were measured under two conditions: normal visibility and fog. Significant adaptation effects in longitudinal driving behavior and a significant increase in mental workload were observed. A new estimation method was used to investigate the extent to which fog influenced the position of so-called action points in the (Δv, s) plane of a psycho-spacing model, where Δv was relative speed and s was spacing. In addition, multivariate regression analysis was applied to investigate the extent to which an influence could be observed on acceleration and on jumps in acceleration at the action points. Large differences in the positions of action points in the (Δv, s) plane, acceleration, and jumps in acceleration were observed between conditions; therefore, car-following patterns closely resemble those predicted by psycho-spacing theory. However, a large degree of inter- and intradriver heterogeneity was observed, possibly caused by differences in mental workload within and between drivers. This heterogeneity indicates that the assumption of deterministic perceptual thresholds is unrealistic and necessitates the development of a data-driven stochastic model based on the principles of psycho-spacing models.

Estimation of the time needed to evacuate a population from a threatened area in case of a disaster is one of the main issues in the design of an evacuation plan. The challenge is to develop a strategy that optimally uses the network capacity to minimize the total evacuation time. In this paper, the impacts of various departure time spans on evacuation time and network performance are investigated with a microscopic traffic simulation model. The network performance has been analyzed with the use of the macroscopic fundamental diagram (MFD). Although the MFD usually shows a decrease in travel production after a peak is reached, this is not the case in the simulation of evacuation scenarios. The outflow of the network remains constant because it depends on the capacity of local bottlenecks upstream of the limited number of destinations, but the number of vehicles in the network increases because of an increase in congestion. Although the overall network performance is insensitive to the evacuation time spans, it is observed that for shorter evacuation time spans, internal gridlock effects cause lengthy delays for specific groups of drivers, who, it turns out, are unable to leave the network in time. The use of a simulation study in combination with an MFD can therefore identify the routes used and the bottlenecks on these routes leading to the destinations, while the maximum production level (defined as the number of arrivals, determined in the MFD) indicates the optimal level of demand.

Various decision theories have been used to explain travelers' behavior. This paper presents a comparative analysis from the points of view of theory and application of the expected utility theory, prospect theory, and regret theory. The application was based on an empirical data set on route choice behavior with and without information provision. Results show that despite the widespread use of expected utility theory to model travelers' behavior, the use of prospect theory is quite appropriate and promising, especially when information is provided. The reference point plays an important role in the prediction ability of prospect theory. The greatest prediction ability occurs when the reference point is aligned with the observed behavior and thus reinforces the necessity of establishing appropriate and meaningful values. This study empirically shows the potential of alternatives to expected utility theory to capture travelers' behavior better, as in the case of prospect theory under the proposed model specification, but this is not necessarily true, as demonstrated by the results obtained by use of regret theory.

A new data-driven stochastic car-following model based on the principles of psychospacing or action-point modeling is presented. It uses empirical or experimental trajectory data and mimics the main microscopic behavioral characteristics present in the data. In the action-point model, regions are defined in the relative speed–distance headway plane, in which the follower is likely to perform an action (increase or decrease acceleration) or not. These regions can be established empirically from vehicle trajectory data and thereby yield a joint cumulative probability distribution function of the action points. Furthermore, the conditional distribution of the actions (the size of the acceleration or deceleration given the current distance headway and relative speed or given the acceleration before the action) can be determined from these data as well. To assess the data correctly, a new filtering technique is proposed. The main hypothesis behind this idea is that the speed profile is a continuous piecewise linear function: accelerations are piecewise constant changing values at nonequidistant discrete time instants. The durations of these constant acceleration periods are not fixed but depend on the state of the follower in relation to its leader. The data analysis illustrates that driving behavior shows nonequidistant constant acceleration periods. The distributions of the action points and the conditional accelerations form the core of the presented data-driven stochastic model. The mathematical formalization that describes how these distributions can be used to simulate car-following behavior is presented. Empirical data collected on a Dutch motorway are used to illustrate the workings of the approach and the simulation results.

One of the key traffic variables required for the ex post and ex ante evaluation of traffic management and policy measures is origin–destination (O-D) demand matrices. Without ground truth O-D information, however, it is difficult, if not impossible, to assess the quality of an O-D estimation method because so many unknowns are involved. One indicator of the quality of an O-D estimation method is the sensitivity of the method to, and its robustness against, random and structural perturbations of the input from a few typical test networks (e.g., data from sensors, prior O-D matrices). In this paper, an assessment methodology is proposed on the basis of the Latin hypercube method, which is an efficient alternative to Monte Carlo sampling and particularly suited for high-dimensional estimation problems. The methodology is demonstrated on a real urban corridor network for a well-known O-D estimation method (the minimum information estimation method) to illustrate the results that can be obtained and how these results can be used to benchmark different O-D estimation methods.

The ability of microscopic (simulation) models to represent lane-changing behavior according to reality has recently been questioned. In this paper the merging maneuver (a specific type of lane changing) is analyzed with empirical data. First, a conceptual model is composed; it includes the factors influencing merging behavior, namely the merge location and its relation to prevailing driving conditions, gap acceptance, and the relaxation phenomenon. The empirical data set consists of 35 min of vehicle maneuvers on 400 m of freeway, collected by a camera mounted underneath a helicopter. This process results in a data set of 3,459 vehicle trajectories, from which 704 trajectories describe merging vehicles. It is found that different merge locations are used under congested and freeflow traffic conditions. During free-flow, most vehicles merge at the first half of the acceleration lane. Under congested traffic conditions, relatively more merges are registered at the end of the acceleration lane. The smallest accepted gap observed in the data set lies between 0.75 and 1.0 s. Net headways between the merging vehicle and the new leader and new follower of less than 0.25 s are recorded. These short accepted gaps are growing over time and indicate relaxation behavior. From the data analysis it can be concluded that gap acceptance theories, as they are used in current models and theories to model merge behavior, are not able to model the observed behavior accurately.

Macroscopic fundamental diagrams (MFDs) exist in large urban networks in which traffic conditions are homogenous. They can be used for estimation of the level of service on road networks, perimeter control, and macroscopic traffic modeling. However, before the MFD concept can be applied, the factors that influence the MFD shape should be identified and their effects investigated. A microscopic simulation model is used to change conditions, that is, to derive MFDs under different conditions and for different types of networks. Results indicate that a relationship indeed exists between production and accumulation for the whole network as well as for parts of the network focused on freeway or urban links. MFD shape is a property not only of the network itself but also of the applied traffic control measures. At the same time, congestion onset and resolution lead to heterogeneous traffic conditions with congestion at specific locations in the network, resulting in loops in congested parts of the MFD. Investigation of the effect of traffic demand on MFD also indicates that rapidly changing traffic demands drastically affect MFD shape.

The values on parameters describing longitudinal driving behavior in car-following models differ substantially between drivers. Different individual interactions with the environment are assumed to play an important role, which might be explained through mental workload. Therefore a driving simulator experiment with a repeated measures design was performed to investigate to what extent perception of an incident in the other driving lane influences physiological indicators as well as subjective estimates of mental workload and longitudinal driving behavior. As almost none of the current models of car-following behavior incorporate mental workload as a determinant of driving behavior, an investigation was conducted by using a calibration approach for joint estimation to determine whether these models, represented by the intelligent driver model and the Helly model, adequately described longitudinal driving behavior in case of incidents in the other driving lane. The results indicated that perception of an incident in the other driving lane influenced mental workload as measured by physiological indicators and longitudinal driving behavior. In addition, the results indicated that current car-following models did not adequately describe driving behavior in case of incidents in the other driving lane.

The kinematic wave model is often used in simulation tools to describe dynamic traffic flow and to estimate and predict traffic states. Discretization of the model is generally based on Eulerian coordinates, which are fixed in space. However, the Lagrangian coordinate system, in which the coordinates move with the velocity of the vehicles, results in more accurate solutions. Furthermore, if the model includes multiple user classes, it describes real traffic more accurately. Such a multiclass model, in contrast to a mixed-class model, treats different types of vehicles (e.g., passenger cars and trucks or vehicles with different origins or destinations, or both) differently. The Lagrangian coordinate system is combined with a multiclass model, and a Lagrangian formulation of the kinematic wave model for multiple user classes is proposed. It is shown that the advantages of the Lagrangian formulation also apply for the multiclass model. Simulations based on the Lagrangian formulation result in more accurate solutions than simulations based on the Eulerian formulation.

Microscopic simulation models have become widely applied tools in traffic engineering. Nevertheless, parameter identification of these models remains a difficult task. This is partially because parameters are generally not directly observable from common traffic data; also there is a lack of reliable statistical estimation techniques. This study puts forward a new general and structured approach to identifying parameters of car-following models. One of the main contributions of this study is joint estimation of parameters for multiple vehicles. Furthermore, prior information on the parameter values (or the valid range of values) can be estimated. The study also deals with serial correlation in the trajectory data. In doing so, the newly developed approach generalizes the maximum likelihood estimation approach proposed by the authors. The approach allows for statistical analysis of the model estimates, including the standard error of the parameter estimates and the correlation of the estimates. With the likelihood ratio test, models of different complexity (defined by the number of model parameters) can be cross-compared. A useful property of this test is that it takes into account the number of parameters of a model as well as the performance. The approach is applied to car-following behavior by using Dutch freeway vehicle trajectories collected from a helicopter.

Dynamic traffic assignment (DTA) models typically describe travelers selecting their routes before departure (pretrip) or during the trip (en route). However, in reality, people follow a certain route but have the opportunity to deviate from that route. An analytical hybrid route choice model is proposed that unifies pretrip and en route route choice in a tractable way. It enables modeling intermediate states where travelers make pretrip route choice decisions and may deviate from this route if they receive information about a more attractive route, for instance, because of unforeseen adverse traffic conditions. The hybrid route choice model is widely applicable to various planning and management applications in DTA and makes the DTA model more realistic in cases such as route guidance problems, where the combination of prescribed routes and en route route choice is evident. Furthermore, the proposed route choice model is generic because different dynamic traffic flow models can be used in the model, analytical or simulation-based. Also, two common problems in DTA related to gridlock and time-varying network conditions are solved in the hybrid route choice model.

Video data are being used more often to study traffic operations. However, extracting vehicle trajectories from video by current methods is a difficult process, typically resulting in many errors. The process requires extensive labor to correct the trajectories manually. This paper proposes a method to process video data from traffic operations. Instead of detecting a vehicle in each picture of the video separately, the video data are transformed so that the trajectories of the vehicles (their position over time) become visible in a single image. In this single image, the trajectories can be found by detecting lines. The difference from other methods is that trajectories rather than vehicles are detected. Trajectory (line) detection is more robust than vehicle (rectangle) detection; with this method, about 95% of the trajectories are detected correctly and, more important, the segments of each trajectory are much longer compared with results from other methods in the literature. Also, the detection is a quick process because only a single image is required to be analyzed. For a data set 5 min long, transforming costs several minutes, and automatically detecting and tracking costs 40 to 50 min per lane. Manual correction is then necessary, which costs about 10 min per lane. In contrast, with a different method the total processing time for analyzing traffic operations costs about 1 week for all lanes together.

Parameter values obtained by calibrating a car-following model using trajectory observations can provide important new insights into longitudinal driving behavior that cannot be derived from double loop detector data. However, that holds only if these behavioral parameters can be reliably estimated. This important aspect of calibration has been generally neglected in scientific research. In this contribution, therefore, calibrations are done to analyze several aspects influencing the degree to which behavioral parameters can be reliably estimated. In these calibrations trajectory data created by the authors with known characteristics as well as empirical trajectories are used. It is shown that traffic conditions during observation determine which model parameters can be reliably estimated. That measurement errors do negatively influence the reliability of parameter estimates is also demonstrated. The same negative effect appears to occur when a car-following model not fully able to capture the dynamics of the observed driver is calibrated. Finally, it is found that the extent to which a specific parameter can be reliably estimated is dependent on the degree to which it affects model predictions and thereby the value of the calibration objective. From this contribution it is concluded that to answer a specific research question on longitudinal driving behavior, it is essential to collect trajectories during appropriate traffic conditions. The results furthermore stress that even when trajectories contain enough information to calibrate the behavioral parameters of interest, whether it is reliable enough to draw inferences on car-following behavior needs to be determined for every parameter value separately.

This investigation focuses on how the heterogeneity of pedestrian characteristics influences the buildup of congestion and affects the efficiency of pedestrian flows. Three commonly used parameters in pedestrian models–-desired speed, body size, and reaction time–-were varied in the population. Real pedestrian flows are heterogeneous regarding pedestrian characteristics. However, not much is known about the way that affects the qualities of the flow and how important it is to the outcomes of microsimulation models. The NOMAD model developed by Delft University of Technology is used to perform simulations in which the aforementioned heterogeneity is introduced. The investigation was carried out by creating bidirectional flows with fixed demands. The flows were analyzed by observing the development of breakdowns, average speeds, and average densities for different demands. It is shown that the influence of heterogeneity on breakdown probabilities and flow efficiency is considerable. To investigate this further, the dynamic lane formation process is investigated in detail. In addition to further insights into the causes for breakdown, it is found that the number of lanes increases with the decrease in heterogeneity in desired speed and in body size. However the opposite happens for heterogeneity in reaction time. Results indicate that heterogeneity in the population has a large impact on the flow quality and should be included in models explicitly to improve prediction performance.

Traffic state estimation plays an important role in operational traffic management and is essential for real-time traffic modeling and prediction. As more heterogeneous traffic data [e.g., from loops, probe vehicles, advanced vehicle identification (AVI) systems] become available, data fusion has become one of the main challenges of state estimation. Given the different semantics over space and time (e.g., AVI data and local data from loops), data fusion is a far more complex problem than it appears at first glance. A new algorithm for fusing data from local detectors (loops) with travel times obtained from AVI systems was developed. This simple but mathematically elegant algorithm— called piecewise inverse speed correction—by using individual travel times (PISCIT) correctly and efficiently combines these data (in essence incompatible) and produces a state estimate (space mean speeds per cell), which is better than one obtained by any of the data sources individually. Moreover, PISCIT is robust with respect to structural and random errors in the source data. The approach is validated using synthetic data generated by microscopic simulation. The algorithm corrects the traffic state correctly even when nonuniform deviations are up to 70%.

The heterogeneity of traffic is a significant if not dominant factor in accurately modeling freeway traffic flow operations. For example, high truck percentages may induce congestion at much lower volumes, and hence different network traffic conditions may result than with low truck percentages. This implies that traffic models for real-time decision support systems in traffic management centers should provide the means to account for traffic heterogeneity. A new, multiclass, first-order traffic model is presented that provides these means and is implemented in the decision-support system BOSS-Offline, operational in all five highway traffic management centers in the Netherlands. FASTLANE differs from earlier multiclass first-order macroscopic traffic models in that it calculates the dynamics in terms of state-dependent (instead of constant) passenger-car equivalents, which is in line with both theory and empirical microscopic data. The model is numerically solved by an efficient and stable Godunov-based solver while maintaining a dynamic and realistic representation of class-specific flows and densities throughout the network. In two synthetic test cases and one based on real data, the workings of FASTLANE under different truck percentages and different conditions are demonstrated.

Incidents on freeways cause large delays for road users. These delays depend largely on the capacity at the incident location, which is determined by the drivers’ behavior at the accident location. Few empirical facts are available on traffic operations during an incident. This paper presents high-quality videos of the traffic flow around two accidents recorded from a helicopter. From the collected images, traffic counts have been performed at the exact location of the incident. This has two advantages. First, the capacity at the bottleneck per lane could be estimated. Second, truck counts could be converted to passenger car units at the location of the bottleneck. Counts show that the (outflow) capacity of the remaining lanes is about 50% lower than the (free-flow) capacity of the same number of lanes. This means that the road capacity in the opposite direction is reduced by half by the rubbernecking effect. The capacity of the road in the direction of the accident is reduced by more than half because not all lanes are in use. The images provide information on the causes for the capacity reduction. A leader accelerates and the follower accelerates a short time later. The average time between these two accelerations is estimated at about 3 s, but the video also shows a large spread of these times. The results can be used to assess consequences of incidents, in an analytical way and in macroscopic or microscopic traffic simulators.

Interest in calibration of car-following models by using real-life microscopic trajectory data is increasing. However, more information is needed on the influence of methodological issues on calibration results as well as on the influence of practical issues related to the use of real-life data. In particular, the influence of measurement errors on parameter estimates has not yet been considered in detail. To gain insight into the influence of measurement errors on calibration results, synthetic data were created to which several types of measurement error are introduced. These data are input to a validated calibration procedure, after which it is studied how well the parameters used for creating the data can be identified from the erroneous data. The sensitivity of the objective function to small changes in the optimal parameters also is assessed. The calibrations are repeated by using different variables in the objective. The three main findings are that (a) measurement errors can yield a considerable bias in the estimation results, (b) parameters minimizing the objective function do not necessarily capture following dynamics best, and (c) measurement errors substantially reduce the sensitivity of the objective function and consequently reduce the reliability of estimation results. The extent to which these problems caused by measurement errors can be avoided by smoothing the data carefully before use is assessed and discussed.

Modeling breakdown probabilities or phase-transition probabilities is an important issue when assessing and predicting the reliability of traffic flow operations. Looking at empirical spatiotemporal patterns, these probabilities clearly are a function not only of the local prevailing traffic conditions (density, speed) but also of time and space. For instance, the probability that a start-stop wave occurs generally increases when moving upstream away from the bottleneck location. A simple partial differential equation is presented that can be used to model the dynamics of breakdown probabilities, in conjunction with the well-known kinematic wave model. The main assumption is that the breakdown probability dynamics satisfy the way information propagates in a traffic flow, that is, they move along with the characteristics. The main result is that the main characteristics of the breakdown probabilities can be reproduced. This is illustrated through two examples: free flow to synchronized flow (F-S transition) and synchronized to jam (S-J transition). It is shown that the probability of an F-S transition increases away from the on ramp in the direction of the flow; the probability of an S-J transition increases as one moves upstream in the synchronized flow area.