Xiqun Chen

A Markov Model for Headway/Spacing Distribution of Road Traffic

In this paper, we link two research directions of road traffic-the mesoscopic headway distribution model and the microscopic vehicle interaction model-together to account for the empirical headway/spacing distributions. A unified car-following model is proposed to simulate different driving scenarios, including traffic on highways and at intersections. Unlike our previous approaches, the parameters of this model are directly estimated from the Next Generation Simulation (NGSIM) Trajectory Data. In this model, empirical headway/spacing distributions are viewed as the outcomes of stochastic car-following behaviors and the reflections of the unconscious and inaccurate perceptions of space and/or time intervals that people may have. This explanation can be viewed as a natural extension of the well-known psychological car-following model (the action point model). Furthermore, the fast simulation speed of this model will benefit transportation planning and surrogate testing of traffic signals.

Multimodel Ensemble for Freeway Traffic State Estimations

Freeway traffic state estimation is a vital component of traffic management and information systems. Macroscopic-model-based traffic state estimation methods are widely used in this field and have gained significant achievements. However, tests show that the inherent randomness of traffic flow and uncertainties in the initial conditions of models, model parameters, and model structures all influence traffic state estimations. To improve the estimation accuracy, this paper presents an ensemble learning framework to appropriately combine estimation results from multiple macroscopic traffic flow models. This framework first assumes that any models existing are imperfect and have their own strengths/weaknesses. It then estimates the online traffic states in a rolling horizon scheme. This framework automatically ensembles the information from each individual estimation model based on their performance during the selected regression horizon. In particular, we discuss three weighting algorithms, namely, least square regression, ridge regression, and lasso, which represent different presumptions of model capabilities. A field test based on real freeway measurements indicates that lasso ensemble best handles various uncertainties and improves estimation accuracy significantly. It should be also pointed out that the proposed framework is a flexible tool to assemble nonmodel-based traffic estimation algorithms. This framework can be also extended for many other applications, including traffic flow prediction and travel-time prediction.

Freeway Travel-Time Estimation Based on Temporal–Spatial Queueing Model

Travel time serves as a fundamental measurement for transportation systems and becomes increasingly important to both drivers and traffic operators. Existing speed interpolation algorithms use the average speed time series collected from upstream and downstream detectors to estimate the travel time of a road link. Such approaches often result in inaccurate estimations or even systematic bias, particularly when the real travel times quickly vary. To get rid of this problem, Coifman proposed a creative interpolation algorithm based on kinetic-wave models. This algorithm reconstructs vehicle trajectories according to the velocities and the headways of vehicles. However, it sometimes gives significant biased estimation, particularly when jams emerge from somewhere between the upstream and downstream detectors. To make an amendment, we design a new algorithm based on the temporal-spatial queueing model to describe the fast travel-time variations using only the speed and headway time series that is measured at upstream and downstream detectors. Numerical studies show that this new interpolation algorithm could better utilize the dynamic traffic flow information that is embedded in the speed/headway time series in some special cases.

Developing a 24-Hour Large-Scale Microscopic Traffic Simulation Model for the Before-and-After Study of a New Tolled Freeway in the Washington, DC-Baltimore Region

AbstractFor determining highly disaggregate details about traffic dynamics, microscopic traffic simulation has long proven to be a valuable tool for the evaluation of development plans and operation/control strategies. With recent advances in computing capabilities, research interest in large-scale microscopic simulation has never been greater. This case study develops a 24-h large-scale microscopic traffic simulation model for the Washington, DC, metropolitan area. The model consists of over 7,000 links, 3,500 nodes, 400 signalized intersections, and over 40,000 origin-destination pairs. Various field measurements, such as time-dependent traffic counts and corridor travel times, have been used for model calibration/validation. The EPA’s Motor Vehicle Emission Simulator is linked with the microscopic simulation model for the estimation of environmental impacts. The calibrated model system has been used to comprehensively evaluate a newly built toll road in Maryland, the Intercounty Connector. Various netw...

Surrogate-Based Optimization for Solving a Mixed Integer Network Design Problem

This paper considers the bilevel mixed network design problem (MNDP) used in dynamic traffic assignment (DTA) and the simulation-based optimization solution. The upper level of the bilevel MNDP minimizes the network cost in terms of average travel time by the expansion of existing links and the addition of new candidate links. The lower level is a dynamic user-optimal condition that can be formulated as a variational inequality problem. The MNDP simultaneously finds optimal capacity expansions of existing links and new link additions. A surrogate-based optimization (SBO) framework is proposed for solving the MNDP that is characterized by expensive-to-evaluate objective functions. Because simulation was applied to evaluate those objective functions, additional complexity arose from the fine-grained representation of traffic dynamics in large-scale networks, which were not fully considered by the traditional static user equilibrium. SBO methods enjoy both the advantages of simulation in time-varying network performance evaluation and the efficiency of mathematical optimization. To be more specific, SBO produces computational time savings by exploring the input–output mapping surface in a more systematic and efficient way. For demonstrative purposes, a case study was conducted on the large-scale Montgomery County network in Maryland. In this example, a mesoscopic simulation-based DTA model, DTALite, was used to evaluate the system performance in response to various network design strategies. Results showed that the optimal investment with a moderate budget could reduce 17.73% of the network average travel time in the morning peak. The proposed framework is a general approach, which is ready for application to either continuous or discrete network design problems.

Phase Diagram Analysis Based on a Temporal-Spatial Queueing Model

In this paper, we propose a simple temporal-spatial queueing model to quantitatively address some typical congestion patterns that were observed around on/off-ramps. In particular, we examine three prime factors that play important roles in ramping traffic scenarios: the time τ_in for a vehicle to join a jam queue, the time τ_out for this vehicle to depart from this jam queue, and the time interval <i>T</i> for the ramping vehicle to merge into the mainline. Based on Newells simplified car-following model, we show how τ_in changes with the main road flow rate <i>q</i>_main. Meanwhile, <i>T</i> is the reciprocal of the ramping road flow rate <i>q</i>_ramp. Thus, we analytically derive the macroscopic phase diagram plotted on the <i>q</i>_main-versus- <i>q</i>_ramp plane and τ_in-versus-<i>T</i> plane based on the proposed model. Further study shows that the new queueing model not only reserves the merits of Newells model on the microscopic level but helps quantify the contributions of these parameters in characterizing macroscopic congestion patterns as well. Previous approaches distinguished phases merely through simulations, but our model could derive analytical boundaries for the phases. The phase transition conditions obtained by this model agree well with simulations and empirical observations. These findings help reveal the origins of some well-known phenomena during traffic congestion.

Short-term Forecasting of Passenger Demand under On-demand Ride Services: A Spatio-temporal Deep Learning Approach

Abstract Short-term passenger demand forecasting is of great importance to the on-demand ride service platform, which can incentivize vacant cars moving from over-supply regions to over-demand regions. The spatial dependencies, temporal dependencies, and exogenous dependencies need to be considered simultaneously, however, which makes short-term passenger demand forecasting challenging. We propose a novel deep learning (DL) approach, named the fusion convolutional long short-term memory network (FCL-Net), to address these three dependencies within one end-to-end learning architecture. The model is stacked and fused by multiple convolutional long short-term memory (LSTM) layers, standard LSTM layers, and convolutional layers. The fusion of convolutional techniques and the LSTM network enables the proposed DL approach to better capture the spatio-temporal characteristics and correlations of explanatory variables. A tailored spatially aggregated random forest is employed to rank the importance of the explanatory variables. The ranking is then used for feature selection. The proposed DL approach is applied to the short-term forecasting of passenger demand under an on-demand ride service platform in Hangzhou, China. The experimental results, validated on the real-world data provided by DiDi Chuxing, show that the FCL-Net achieves the better predictive performance than traditional approaches including both classical time-series prediction models and state-of-art machine learning algorithms (e.g., artificial neural network, XGBoost, LSTM and CNN). Furthermore, the consideration of exogenous variables in addition to the passenger demand itself, such as the travel time rate, time-of-day, day-of-week, and weather conditions, is proven to be promising, since they reduce the root mean squared error (RMSE) by 48.3%. It is also interesting to find that the feature selection reduces 24.4% in the training time and leads to only the 1.8% loss in the forecasting accuracy measured by RMSE in the proposed model. This paper is one of the first DL studies to forecast the short-term passenger demand of an on-demand ride service platform by examining the spatio-temporal correlations.

Characterising scattering features in flow–density plots using a stochastic platoon model

The scattering features of points in flow–density plot remain as an attractive topic in the last several decades. Some previous studies either assumed that the points of congested traffic flows were completely random or that the implicit rules of hidden distribution were difficult to describe. Although the scattering features are influenced by various factors (e.g. lane-changing manoeuvers, merging behaviours and driver heterogeneity), we believe that they are mainly dominated by the microscopic headway/spacing distributions. In this paper, we relax the assumption of deterministic headway/spacing in Newells simplified car-following model and allow random headways/spacings in a homogeneous platoon (vehicles run closely at the same velocity). Further extending the conventional deterministic reciprocal relationship between flow rate and headway, we find that the reciprocal of average headway of a homogeneous platoon and the corresponding flow rate should follow the same distribution. Based on these two extensions, we can link the conditional distributions of average headway in a homogeneous platoon and the conditional distributions of flow rate, all with respect to velocity. When the aggregation time interval is small enough (e.g. 30 s), tests on Performance Measurement System (PeMS) data reveal that the seemingly disorderly scattering points in the macroscopic flow–density plot follow the estimated flow rate distributions from Next Generation Simulation vehicular trajectories. While if the aggregation time interval increases (e.g. to 5 min), the measured vehicles probably pass the loop detectors at different velocities and form heterogeneous platoons. It becomes difficult to find a definite distribution model that can fit average headway/spacing for heterogeneous platoons. However, most points in flow–velocity plot still locate within a certain 2D region, whose boundaries can be obtained from the homogeneous platoon model. Finally, tests on PeMS data verify the estimated boundaries.

Location Specific Cell Transmission Model for Freeway Traffic

Abstract This paper describes a location specific cell transmission model of freeway traffic based on the observed variability of fundamental diagrams both along and across freeway segments. This model extends the original cell transmission model (CTM) mechanism by defining various shapes of fundamental diagrams to reproduce more complex traffic phenomena, including capacity drops, lane-by-lane variations, nonhomogeneous wave propagation velocities, and temporal lags. A field test on a Canadian freeway was used to demonstrate the validity of the location specific CTM. The simulated spatio-temporal evolutions of traffic flow show that the model can be used to describe the traffic dynamics near bottlenecks more precisely than the original model.

Social Welfare Maximization of Multimodal Transportation: Theory, Metamodel, and Application to Tianjin Ecocity, China

Multimodal urban transportation systems exhibit complex inter actions between components, including users, multimodal transportation facilities, supply side agencies, and operators. Although these interactions are obvious, rigorous quantitative methods for optimizing control variables across modes of transportation on real-world networks are deficient. A social welfare maximization model was established for joint optimization of bus fare, rail transit fare, and congestion tolls for private cars. The authors determined the optimality conditions and the second-order partial derivatives for this optimization problem. Because of the complexity of the multimodal urban transportation system, the objective function of social welfare has no closed form and is extremely expensive to evaluate. The authors therefore proposed a simulation-based framework for evaluation of the objective function and optimization of three decision variables across multiple travel modes. Several metamodels (i.e., mathematical functions that approximate the true shape of an unknown, nonlinear, and complex objective function) were adopted to approximate the highly nonlinear input–output mappings in the urban system. This is the first study to develop a simulation-based method for joint optimization of transit and road network operations. The case study applied the simulation-based optimization framework to the Sino–Singapore ecocity in Tianjin, China, by using VISSUM as the urban systems simulator. An “ecocity” is defined by the authors as “a thriving city that is socially harmonious, environmentally friendly, resource-efficient, and a model for sustainable development.” Results show that metamodels can accurately approximate the real objective function and produce good suboptimal and near-optimal solutions. The optimal combination of transit fares and congestion tolls significantly outperform those under two baseline scenarios. The optimal solutions also suggest that extreme transit fares (too high or too low) or congestion tolls are contrary to welfare-maximizing objectives. The proposed method can be applied for joint optimization of other multimodal planning and operational strategies, such as investment and operational decisions across various modes of transportation.