Integrated model for traffic flow forecasting under rainy conditions

Inclement weather, such as heavy rain, significantly affects road traffic flow operation, which may cause severe congestion in road networks in cities. This study investigates the effect of inclement weather, such as rain events, on traffic flow and proposes an integrated model for traffic flow parameter forecasting during such events. First, an analysis of historical observation data indicates that the forecasting error of traffic flow volume has a significant linear correlation with mean precipitation, and thus, forecasting accuracy can be considerably improved by applying this linear correlation to correct forecasting values. An integrated online precipitation‐correction model was proposed for traffic flow volume forecasting based on these findings. We preprocessed precipitation data transformation and used outlier detection techniques to improve the efficiency of the model. Finally, an integrated forecasting model was designed through data fusion methods based on the four basic forecasting models and the proposed online precipitation‐correction model. Results of the model validation with the field data set show that the designed model is better than the other models in terms of overall accuracy throughout the day and under precipitation. However, the designed model is not always ideal under heavy rain conditions. Copyright © 2016 John Wiley & Sons, Ltd.     

Predicting short-term bus passenger demand using a pattern hybrid approach

This paper proposes an Interactive Multiple Model-based Pattern Hybrid (IMMPH) approach to predict short-term passenger demand. The approach maximizes the effective information content by assembling the knowledge from pattern models using historical data and optimizing the interaction between them using real-time observations. It can dynamically estimate the priori pattern models combination in advance for the next time interval. The source demand data were collected by Smart Card system along one bus service route over one year. After correlation analysis, three temporal relevant pattern time series are generated, namely, the weekly, daily and hourly pattern time series. Then statistical pattern models are developed to capture different time series patterns. Finally, an amended IMM algorithm is applied to dynamically combine the pattern models estimations to output the final demand prediction. The proposed IMMPH model is validated by comparing with statistical methods and an artificial neural network based hybrid model. The results suggest that the IMMPH model provides a better forecast performance than its alternatives, including prediction accuracy, robustness, explanatory power and model complexity. The proposed approach can be potentially extended to other short-term time series forecast applications as well, such as traffic flow forecast.     

Short-term speed predictions exploiting big data on large urban road networks

Big data from floating cars supply a frequent, ubiquitous sampling of traffic conditions on the road network and provide great opportunities for enhanced short-term traffic predictions based on real-time information on the whole network. Two network-based machine learning models, a Bayesian network and a neural network, are formulated with a double star framework that reflects time and space correlation among traffic variables and because of its modular structure is suitable for an automatic implementation on large road networks. Among different mono-dimensional time-series models, a seasonal autoregressive moving average model (SARMA) is selected for comparison. The time-series model is also used in a hybrid modeling framework to provide the Bayesian network with an a priori estimation of the predicted speed, which is then corrected exploiting the information collected on other links. A large floating car data set on a sub-area of the road network of Rome is used for validation. To account for the variable accuracy of the speed estimated from floating car data, a new error indicator is introduced that relates accuracy of prediction to accuracy of measure. Validation results highlighted that the spatial architecture of the Bayesian network is advantageous in standard conditions, where a priori knowledge is more significant, while mono-dimensional time series revealed to be more valuable in the few cases of non-recurrent congestion conditions observed in the data set. The results obtained suggested introducing a supervisor framework that selects the most suitable prediction depending on the detected traffic regimes.     

Forecasting light-duty vehicle demand using alternative-specific constants for endogeneity correction versus calibration

We investigate parameter recovery and forecast accuracy implications of incorporating alternative-specific constants (ASCs) in the utility functions of vehicle choice models. We compare two methods of incorporating ASCs: (1) a maximum likelihood estimator that computes ASCs post-hoc as calibration constants (MLE-C) and (2) a generalized method of moments estimator that uses instrumental variables (GMM-IV) to correct for price endogeneity. In a synthetic study we observe significant coefficient bias with MLE-C when the price-ASC correlation (endogeneity) is large. GMM-IV successfully mitigates this bias given valid instruments but exacerbates the bias given invalid instruments. Despite greater coefficient bias, MLE-C yields better forecasts than GMM-IV with valid instruments in most of the cases examined, including most cases where the price-ASC correlation present in the estimation data is absent in the prediction data. In a market study of U.S. midsize sedan sales from 2002 – 2006 the GMM-IV model predicts the 1-year-forward market better, but the MLE-C model predicts the 5-year-forward market better. Including an ASC in predictions by any of the methods proposed improves share forecasts, and assuming that the ASC of each new vehicle matches that of its closest competitor vehicle yields the best long term forecasts. We find evidence that the instruments most frequently used in the automotive demand literature may be invalid.     

Applications of Random-Utility-based Multi-region Input–Output Models of Transport and the Spatial Economy

Abstract

An introduction to random-utility-based multiregional input–output models used for the purpose of spatial economic and transport interaction modelling is provided. The main methodological developments and important results of a dozen applications from the years 1996–2013 are described. This is followed by an outlook of potential future directions. Further research is mainly needed in five areas: (a) overall validation of the method, perhaps through back-casting applications on infrastructure plans with observed trade impacts; (b) extensions of trade coefficient models to add realism and improve accuracy; (c) the use of multi-scale modelling to capture interdependencies between geographical scales and to improve the representation of exports and imports; (d) improvements in the representation of price effects, as well as innovation and technological progress, by way of variable technical coefficients; and (e) a deeper investigation of the algorithm used to include elastic selling prices.     

Tests of assignment accuracy: An interurban case study

This paper is concerned with methods of testing the accuracy of traffic assignments. It focuses on the fact that whereas assignment models are usually based on a behavioural hypothesis about drivers' route choice (e.g. cost or time minimisation) the test of the accuracy of the assignment is the extent to which observed link loadings are reproduced. This inconsistency opens up the doubt that an apparently “accurate” assignment on this basis may be a result of compensating errors. It is difficult to apply the same test to accuracy of route choices as is applied to accuracy of link loadings (e.g. chi square, correlation coefficient) and hence a new measure is devised here which can be applied both to comparisons between observed and predicted route choices and comparisons between observed and predicted loadings. It is, moreover, possible to devise a test of significance for this measure so that one can test whether a predicted assignment is significantly different from what one, might have observed on the basis of chance observation. A case study is carried out to test the proposed method. Traffic flows between 72 origins and destinations on either side of the Pennine Mountains in Britain are assigned to a network using different assignment techniques with varied parameters. In all, one hundred and ninety assignments are carried out and the degree of correspondence between observed and predicted route choices and link loadings is measured. The results tend to confirm that the link loadings criterion is not a very stable criterion and that the route choice correspondence criteria seems to behave in a sensible way. A simulation exercise is carried out which produces the probability distribution of the “route-fit” index for different assumed sample levels. The paper concludes by suggesting avenues for further research.     

Classification modeling approach for vehicle dynamics model‐integrated traffic simulation assessing surrogate safety

To assess safety impacts of untried traffic control strategies, an earlier study developed a vehicle dynamics model‐integrated (i.e., VISSIM‐CarSim‐SSAM) simulation approach and evaluated its performance using surrogate safety measures. Although the study found that the integrated simulation approach was a superior alternative to existing approaches in assessing surrogate safety, the computation time required for the implementation of the integrated simulation approach prevents it from using it in practice. Thus, this study developed and evaluated two types of models that could replace the integrated simulation approach with much faster computation time, feasible for real‐time implementation. The two models are as follows: (i) a statistical model (i.e., logit model) and (ii) a nonparametric approach (i.e., artificial neural network). The logit model and the neural network model were developed and trained on the basis of three simulation data sets obtained from the VISSIM‐CarSim‐SSAM integrated simulation approach, and their performances were compared in terms of the prediction accuracy. These two models were evaluated using six new simulation data sets. The results indicated that the neural network approach showing 97.7% prediction accuracy was superior to the logit model with 85.9% prediction accuracy. In addition, the correlation analysis results between the traffic conflicts obtained from the neural network approach and the actual traffic crash data collected in the field indicated a statistically significant relationship (i.e., 0.68 correlation coefficient) between them. This correlation strength is higher than that of the VISSIM only (i.e., the state of practice) simulation approach. The study results indicated that the neural network approach is not only a time‐efficient way to implementing the VISSIM‐CarSim‐SSAM integrated simulation but also a superior alternative in assessing surrogate safety. Copyright © 2014 John Wiley & Sons, Ltd.     

Bus journey times in medium size urban areas

The predictive accuracy of the models based on the fundamental relation between journey time and passenger demand can be improved through data disaggregation or route segmentation. Primary reason for this is the improvement in the estimates of stopping delays and delays due to passenger boarding and/or alighting (dwell time). Both Poisson and Negative Binomial model estimates of stoppings for passenger boarding and alighting are shown to improve with disaggregation. These improvements, however, contribute little to the overall predictability of the fundamental models which are useful for gaining insight into the significance and variability of the stopping delays and dwell time, or testing sensitivity to changes in the long term. Site or route specific models of journey times which have better predictive capability exist, and may be used for short-run planning. However, the interchangeability and performance over time of the latter, have to be evaluated before making definitive conclusions.     

Walking accessibility and property prices

Walking is an imperative travel mode, especially for short trips. Walking accessibility, which is defined as the ease of reaching essential destinations in the walk-in catchment area, may affect property prices because residents are more likely to be willing to pay for this attribute. In addition, different categories of public services may have varied influencing directions and magnitude. These two hypotheses are tested in this study. Taking Xiamen, China as a case study, we estimate the cumulative opportunities of public services on foot and develop a set of hedonic pricing models (more specifically, two pre-specified ordinary least squares models, four Box-Cox transformed models, and two spatial econometric models) to estimate, whether and to what extent, walking accessibility contributes to price premiums (or discounts). Using a database of 22,586 second-hand residential properties in 358 multi- or high-storey residential complexes, we find that (1) walking accessibility to public services contributes to the variations in housing prices and plays a role in determining housing prices; (2) different categories of services have vastly divergent, even opposite, influencing impacts; and (3) walking accessibility to primary schools, commercial centers, and sports and cultural centers have positive effects on house prices whereas walking accessibility to comprehensive hospitals adversely affects housing prices. Methodologically, we confirm that spatial econometric methods improve estimation accuracy and have more explanatory power relative to the standard non-spatial models. Robustness check analysis further guarantees the plausibility of this study.     

Trip misreporting forecast using count data model in a GPS enhanced travel survey

As Global Positioning System (GPS) technology advances, it has been increasingly used to supplement traditional self-reported travel surveys due to its promising features in capturing travel data with better accuracy and reliability. Realizing the limitations of diary-based surveys, this paper presents a study that directly accounts for trip misreporting behavior in trip generation models. Travel data were obtained from prompted-recall assisted GPS survey along with a diary-based survey. Negative Binomial models for count data were developed to accommodate misreporting behavior by introducing interaction effects of the sample-indicator variable with various personal and household variables. The interaction effects indicate how the impacts of the socioeconomic and demographic variables on trip-making vary across the two samples. Assuming that the GPS sample represents the ground truth, the interaction effects actually capture the likelihood and the extent of trip misreporting by detailed personal and household characteristics. The model results reveal significant interaction effects of several personal and household variables, indicating misreporting behavior associated with these attributes. The addition of interaction coefficients to the main effect model represents the real impacts of the independent variables, after compensating for trip misreporting behavior, if any.     

On the role of bridges as anchor points in route choice modeling

This work builds upon the thought that individuals allocate higher levels of importance to some particular features of the route, so called anchor points. Previous route choice models have either ignored the effects of anchor points (route-based models), or have given an exclusive attention to their effects and ignored the behavioral accuracy and practicality of these models (anchor-based models). In this work we argue that the consideration of both route-level attributes and anchor points would enhance the behavioral aspect of route choice models as well as their estimation and prediction abilities. Global Positioning System traces have been used to investigate the effect of bridges as anchor points for trips between Montreal and its Northern suburb, Laval. A classic Nested Logit and a nested Logit Kernel model have been estimated, in which interdependencies among routes crossing the same bridge are captured through the nested structure and the adopted factor analytic approach, respectively. A Metropolis–Hastings path-sampling algorithm is applied, for the first time, on a large road network with more than 40,000 nodes and 19,000 links to provide the consideration choice set. Estimates are then compared to three alternate models, representing route-based and anchor-based formulations; namely Path-Size Logit, Extended Path-Size Logit, and Independent Availability Logit models. Empirical results showed that the proposed nested structures with MH sampling provide better estimates and also perform better in the validation step with respect to comparative models. Findings underscore the importance of considering anchor points in conjunction with route level attributes in route choice decisions.     

Towards data-driven car-following models

Car following models have been studied with many diverse approaches for decades. Nowadays, technological advances have significantly improved our traffic data collection capabilities. Conventional car following models rely on mathematical formulas and are derived from traffic flow theory; a property that often makes them more restrictive. On the other hand, data-driven approaches are more flexible and allow the incorporation of additional information to the model; however, they may not provide as much insight into traffic flow theory as the traditional models. In this research, an innovative methodological framework based on a data-driven approach is proposed for the estimation of car-following models, suitable for incorporation into microscopic traffic simulation models. An existing technique, i.e. locally weighted regression (loess), is defined through an optimization problem and is employed in a novel way. The proposed methodology is demonstrated using data collected from a sequence of instrumented vehicles in Naples, Italy. Gipps’ model, one of the most extensively used car-following models, is calibrated against the same data and used as a reference benchmark. Optimization issues are raised in both cases. The obtained results suggest that data-driven car-following models could be a promising research direction.     

When is big data big enough? Implications of using GPS-based surveys for travel demand analysis

A number of studies in the last decade have argued that Global Positioning Systems (GPS) based survey offer the potential to replace traditional travel diary surveys. GPS-based surveys impose lower respondent burden, offer greater spatiotemporal precision and incur fewer monetary costs. However, GPS-based surveys do not collect certain key inputs required for the estimation of travel demand models, such as the travel mode(s) taken or the trip purpose, relying instead on data-processing procedures to infer this information. This study assesses the impact that errors in inference can have on travel demand models estimated using data from GPS-based surveys and proposes ways in which these errors can be controlled for during both data collection and model estimation. We use simulated datasets to compare performance across different sample sizes, inference accuracies, model complexities and estimation methods. Findings from the simulated datasets are corroborated with real data collected from individuals living in the San Francisco Bay Area, United States. Results indicate that the benefits of using GPS-based surveys will vary significantly, depending upon the sample size of the data, the accuracy of the inference algorithm and the desired complexity of the travel demand model specification. In many cases, gains in the volume of data that can potentially be retrieved using GPS devices are found to be offset by the loss in quality caused by inaccuracies in inference. This study makes the argument that passively collected GPS-based surveys may never entirely replace surveys that require active interaction with study participants.     

Daily activity pattern recognition by using support vector machines with multiple classes

The focus of this paper is to learn the daily activity engagement patterns of travelers using Support Vector Machines (SVMs), a modeling approach that is widely used in Artificial intelligence and Machine Learning. It is postulated that an individual’s choice of activities depends not only on socio-demographic characteristics but also on previous activities of individual on the same day. In the paper, Markov Chain models are used to study the sequential choice of activities. The dependencies among activity type, activity sequence and socio-demographic data are captured by employing hidden Markov models. In order to learn model parameters, we use sequential multinomial logit models (MNL) and multiclass Support Vector Machines (K-SVM) with two different dependency structures. In the first dependency structure, it is assumed that type of activity at time ‘t’ depends on the last previous activity and socio-demographic data, whereas in the second structure we assume that activity selection at time ‘t’ depends on all of the individual’s previous activity types on the same day and socio-demographic characteristics. The models are applied to data drawn from a set of California households and a comparison of the accuracy of estimation of activity types and their sequence in the agenda, indicates the superiority of K-SVM models over MNL. Additionally, we show that accuracy in estimating activity patterns increases using different sets of explanatory variables or tuning parameters of the kernel function in K-SVM.     

Modelling Externalities using Dynamic Traffic Assignment Models: A Review

Recently, there has been a growing interest in externalities in our society, mainly in the context of climate and air quality, which are of importance when policy decisions are made. For the assessment of externalities in transport, often the output of static traffic assignment models is used in combination with so-called effect models. Due to the rapidly increasing possibilities of using dynamic traffic assignment (DTA) models for large-scale transportation networks and the application of traffic measures, already several models have been developed to assess the externalities using DTA models more precisely. Different research projects have shown that there is a proven relation between the traffic dynamics and externalities, such as emissions of pollutants and traffic safety. This means that the assessment of external effects can be improved by using temporal information about flow, speed and density, which is the output of DTA models. In this paper, the modelling of traffic safety, emissions and noise in conjunction with DTA models is reviewed based on an extensive literature survey. This review shows that there are still gaps in knowledge in assessing traffic safety, much research is available concerning emissions, and although little research has been conducted concerning the assessment of noise using DTA models, the methods available can be used to assess the effects. Most research so far has focused on the use of microscopic models, while mesoscopic or macroscopic models may have a high potential for improving the assessment of these effects for larger networks.     

A mixed logit model for predicting exit choice during building evacuations

Knowledge on human behaviour in emergency is crucial to increase the safety of buildings and transportation systems. Decision making during evacuations implies different choices, of which one of the most important concerns is the escape route. The choice of a route may involve local decisions on alternative exits from an enclosed environment. This study investigates the effect of environmental (presence of smoke, emergency lighting and distance of exit) and social factors (interaction with evacuees close to the exits and with those near the decision-maker) on local exit choice. This goal is pursued using an online stated preference survey carried out making use of non-immersive virtual reality. A sample of 1503 participants is obtained and a mixed logit model is calibrated using these data. The model shows that the presence of smoke, emergency lighting, distance of exit, number of evacuees near the exits and the decision-maker and flow of evacuees through the exits significantly affect local exit choice. Moreover, the model indicates that decision making is affected by a high degree of behavioural uncertainty. Our findings support the improvement of evacuation models and the accuracy of their results, which can assist in designing and managing building and transportation systems. The main aim of this study is to enrich the understanding of how local exit choices are made and how behavioural uncertainty affects these choices.     

Three-dimensional conflict count models for unstructured and layered airspace designs

This paper presents analytical models that describe the safety of unstructured and layered en route airspace designs. Here, ‘unstructured airspace’ refers to airspace designs that offer operators complete freedom in path planning, whereas ‘layered airspace’ refers to airspace concepts that utilize heading-altitude rules to vertically separate cruising aircraft based on their travel directions. With a focus on the intrinsic safety provided by an airspace design, the models compute instantaneous conflict counts as a function of traffic demand and airspace design parameters, such as traffic separation requirements and the permitted heading range per flight level. While previous studies have focused primarily on conflicts between cruising aircraft, the models presented here also take into account conflicts involving climbing and descending traffic. Fast-time simulation experiments used to validate the modeling approach indicate that the models estimate instantaneous conflict counts with high accuracy for both airspace designs. The simulation results also show that climbing and descending traffic caused the majority of conflicts for layered airspaces with a narrow heading range per flight level, highlighting the importance of including all aircraft flight phases for a comprehensive safety analysis. Because such trends could be accurately predicted by the three-dimensional models derived here, these analytical models can be used as tools for airspace design applications as they provide a detailed understanding of the relationships between the parameters that influence the safety of unstructured and layered airspace designs.     

A hybrid deep learning based traffic flow prediction method and its understanding

Deep neural networks (DNNs) have recently demonstrated the capability to predict traffic flow with big data. While existing DNN models can provide better performance than shallow models, it is still an open issue of making full use of spatial-temporal characteristics of the traffic flow to improve their performance. In addition, our understanding of them on traffic data remains limited. This paper proposes a DNN based traffic flow prediction model (DNN-BTF) to improve the prediction accuracy. The DNN-BTF model makes full use of weekly/daily periodicity and spatial-temporal characteristics of traffic flow. Inspired by recent work in machine learning, an attention based model was introduced that automatically learns to determine the importance of past traffic flow. The convolutional neural network was also used to mine the spatial features and the recurrent neural network to mine the temporal features of traffic flow. We also showed through visualization how DNN-BTF model understands traffic flow data and presents a challenge to conventional thinking about neural networks in the transportation field that neural networks is purely a “black-box” model. Data from open-access database PeMS was used to validate the proposed DNN-BTF model on a long-term horizon prediction task. Experimental results demonstrated that our method outperforms the state-of-the-art approaches.     

Principal component analysis of urban traffic characteristics and meteorological data

Principal component analysis (PCA) is used to analyze one-year traffic, emission and meteorological data for an urban intersection in the Delhi. The 1997 data include meteorological, traffic and emission variables. In urban intersections the complexities of site, traffic and meteorological characteristic may result in a high cross correlation among the variables. In such situations, PCA can provide an independent linear combination of the variables. Here it is used to analyze 1, 8 and 24 h average emission, traffic and meteorological data. It shows that four principal components for the 24 h average have the highest loadings for traffic and emission variables with a strong correlation between them. PC loadings for the 1 and 8 h data indicate the least variation among them.     

Trajectory data reconstruction and simulation-based validation against macroscopic traffic patterns

This paper shows that the behavior of driver models, either individually or entangled in stochastic traffic simulation, is affected by the accuracy of empirical vehicle trajectories. To this aim, a “traffic-informed” methodology is proposed to restore physical and platoon integrity of trajectories in a finite time–space domain, and it is applied to one NGSIM I80 dataset. However, as the actual trajectories are unknown, it is not possible to verify directly whether the reconstructed trajectories are really “nearer” to the actual unknowns than the original measurements. Therefore, a simulation-based validation framework is proposed, that is also able to verify indirectly the efficacy of the reconstruction methodology. The framework exploits the main feature of NGSIM-like data that is the concurrent view of individual driving behaviors and emerging macroscopic traffic patterns. It allows showing that, at the scale of individual models, the accuracy of trajectories affects the distribution and the correlation structure of lane-changing model parameters (i.e. drivers heterogeneity), while it has very little impact on car-following calibration. At the scale of traffic simulation, when models interact in trace-driven simulation of the I80 scenario (multi-lane heterogeneous traffic), their ability to reproduce the observed macroscopic congested patterns is sensibly higher when model parameters from reconstructed trajectories are applied. These results are mainly due to lane changing, and are also the sought indirect validation of the proposed data reconstruction methodology.