Peter J Jin

Perspectives on Future Transportation Research: Impact of Intelligent Transportation System Technologies on Next-Generation Transportation Modeling

In this paper, we attempt to summarize the impact of technologies, especially intelligent transportation system (ITS) technologies, on transportation research during the last several decades and provide perspectives on how future transportation research may be affected by the availability and development of new ITS technologies. The intended audience of the paper includes young transportation researchers and professionals. Current transportation models are divided into “generations” based on their technological and practical background. Based on the trends in the past and the potential technologies to be implemented in the future, general characteristics of the next generations of transportation models are proposed and discussed to provide a vision regarding expected future achievements in transportation research. This paper is intended to be a working document, in the sense that it will be updated periodically.

Short-Term Traffic Prediction Based on Dynamic Tensor Completion

Short-term traffic prediction plays a critical role in many important applications of intelligent transportation systems such as traffic congestion control and smart routing, and numerous methods have been proposed to address this issue in the literature. However, most, if not all, of them suffer from the inability to fully use the rich information in traffic data. In this paper, we present a novel short-term traffic flow prediction approach based on dynamic tensor completion (DTC), in which the traffic data are represented as a dynamic tensor pattern, which is able capture more information of traffic flow than traditional methods, namely, temporal variabilities, spatial characteristics, and multimode periodicity. A DTC algorithm is designed to use the multimode information to forecast traffic flow with a low-rank constraint. The proposed method is evaluated on real-world data sets and compared with other state-of-the-art methods, and the efficacy of the proposed approach is validated on the experiments of traffic flow prediction, particularly when dealing with incomplete traffic data.

Origin-Destination Estimation for Non-Commuting Trips Using Location-Based Social Networking Data

The development of sustainable mobility solutions calls for significant advances in travel demand data collection beyond the long-term static planning data usually available at planning agencies. This paper proposes a combined clustering, regression, and gravity model to estimate an origin-destination (OD) matrix for non-commuting trips based on Foursquare user check-in data in the Chicago urban area. The estimated OD matrix is found to be similar to the ground-truth OD matrix obtained from CMAP (Chicago Metropolitan Agency for Planning). The potential applications for generating day-of-the-week and dynamic bihourly OD patterns from Foursquare data are also illustrated.

Location-Based Social Networking Data: Exploration into Use of Doubly Constrained Gravity Model for Origin–Destination Estimation

Trip distribution is an invaluable portion of the transportation planning process; this distribution leads to the creation of origin–destination (O-D) matrices. Location-based social networking (LBSN) has increased in popularity and sophistication and has emerged as a new travel demand data source. Users of LBSN provide location-sensitive data interactively with mobile devices, including smartphones and tablets. These data can provide O-D estimates with significantly higher temporal resolution at a much lower cost in comparison with traditional methods. An LBSN O-D estimation model based on the doubly constrained gravity model was proposed to improve a previously proposed model based on the singly constrained gravity model. The proposed methodology was calibrated and comparatively evaluated against the O-D matrix generated by the method based on the singly constrained gravity model as well as a reference matrix from the local metropolitan planning organization. The results of this method illustrate significant improvement in reducing the O-D estimation errors caused by the sampling bias from the method based on the singly constrained gravity model.

Reducing the Error Accumulation in Car-Following Models Calibrated With Vehicle Trajectory Data

With the development of probe vehicle technologies and the emerging connected vehicle technologies, applications and models using trajectory data for calibration and validation significantly increase. However, the error accumulation issue accompanied by the calibration process has not been fully investigated and addressed. This paper explores the mechanism and countermeasures of the error accumulation problems of car-following models calibrated with microscopic vehicle trajectory data. In this paper, we first derive the error dynamic model based on an acceleration-based generic car-following model formulation. The stability conditions for the error dynamic model are found to be different from the model stability conditions. Therefore, adjusting feasible ranges of model parameters in the car-following model calibration to ensure model stability cannot guarantee the error stability. However, directly enforcing those error stability conditions can be ineffective, particularly when explicit formulations are difficult to obtain. To overcome this issue, we propose several countermeasures that incorporate error accumulation indicators into the error measures used in the calibration. Numerical experiments are conducted to compare the traditional and the proposed error measures through the calibration of five representative car-following models, i.e., General Motors, Bando, Gipps, FREeway SIMulation (FRESIM), and intelligent driver model (IDM) models, using field trajectory data. The results indicate that the weighted location mean absolute error (MAE) and the location MAE with crash rate penalty can achieve the best overall error accumulation performance for all five models. Meanwhile, traditional error measures, velocity MAE, and velocity Theils U also achieve satisfactory error accumulation performance for FRESIM and IDM models, respectively.

Estimating Missing Traffic Volume Using Low Multilinear Rank Tensor Completion

Traffic volume data have been collected and used for various purposes in some aspects of intelligent transportation systems (ITS) applications. However, the unavoidable detector malfunction can cause data to be missing. It is often necessary to develop an effective approach to recover the missing data. In most previous methods, temporal correlation is explored to reconstruct missing traffic volume. In this article, a new missing traffic volume estimation approach based on tensor completion is proposed by exploring traffic spatial–temporal information. The tensor model is utilized to represent traffic volume, which allows for exploring the multicorrelation of traffic volume in spatial and temporal information simultaneously. In order to estimate the missing traffic volume represented by the tensor model, a novel tensor completion algorithm, called low multilinear rank tensor completion, is proposed to reconstruct the missing entries. The proposed approach is evaluated on the PeMS database. Experimental results demonstrate that the proposed method is more effective than the state-of-art methods, especially when the ratio of missing data is high.

Determining Strategic Locations for Environmental Sensor Stations with Weather-Related Crash Data

Adverse weather leads to more than 1.5 million vehicular crashes, resulting in 800,000 injuries and 7,000 fatalities nationally. The appropriate deployment of roadway weather information systems (RWIS) has been considered an effective strategy to address safety concerns. However, the current practice of selecting the locations of RWIS stations depends primarily on the knowledge and experiences of field operators. Limited research has been conducted on methodologies that can identify RWIS locations systematically on the basis of widely available safety and geographic information system (GIS) data. This paper proposes a spatial optimization method to identify strategic locations for deploying RWIS stations within a large regional transportation network. Weather-related crash data were converted to a safety concern index and then used to examine routes that provided good spatial coverage of the region for optimal locations for RWIS stations through a maximization algorithm. The proposed method is evaluated with crash and GIS data from a tri-county region in Texas with the coverage for each RWIS station assumed to be 10 mi. The resulting locations illustrate the promising potential for the proposed RWIS location optimization algorithm. Additional sensitivity analysis is also conducted with an evaluation of the resulting changes from yearly crash data variations.

A vehicle type-dependent visual imaging model for analysing the heterogeneous car-following dynamics

Heterogeneity is an essential characteristic in car-following behaviours, which can be defined as the differences between the car-following behaviours of driver/vehicle combination under comparable conditions. This paper proposes a visual imaging model (VIM) with relaxed assumption on (1) a drivers perfect perception for the states of the neighbouring vehicles (e.g. spacing, velocity, etc.) and (2) uniform reaction to vehicles with different sizes in most existing car-following models. VIM utilises the visual imaging information subtended by the preceding vehicle as the stimuli drivers react to, and can generate greater stimuli from the preceding vehicle with larger apparent size (i.e. vehicle width × vehicle height) under short gap distance with the follower, but less change in stimuli from the distant leading vehicle under various apparent sizes. The NGSIM data containing vehicle type/size information is used to evaluate VIM at different levels. At the level of single trajectory pair, the calibrated VIM occupies the well capability of reproducing the trajectory of the follower, and can also reproduce statistical results from the field data, that is, the gap distance for car-following truck (C-T) is greater than that for car-following car (C-C). At the level of vehicle type, the calibration results also show the promising performance of VIM in describing heterogeneous car-following behaviours with the simple model formulation and limited model parameters compared with other six reference models.

Freeway recurrent bottleneck identification algorithms considering detector data quality issues

Computer algorithms used to identify recurrent freeway bottlenecks have been studied since the deployment of loop detecting systems. Such algorithms automatically analyze the archived loop detector data and identify potential recurrent bottlenecks and their characteristics, such as location, time of day, and activation rate, for further investigation. In a highway congestion mitigation project, such algorithms can save time and resources for the initial screening of bottlenecks over a large freeway network. These algorithms include rule-based, contour-map-based, and simulation-based methods. However, existing methods require loop detector data with high accuracy and consistency, which is difficult to achieve in prevailing loop detecting systems. This paper proposes a new bottleneck identification algorithm with strong error and noise tolerance. Several simple denoising methods to improve the error resistance of existing algorithms are also proposed. Using statistical error analysis methods, the proposed algorithm and the denoising methods were calibrated and evaluated using field data collected from two distinct freeway corridors (US 12/14 and I-894) in the U.S. state of Wisconsin. Ground truth data for this study come from the manual inspection of 287,055 traffic video snapshots in the course of a month. In the evaluation tests, the proposed algorithm can produce quality congestion identification results with fewer false alarms than the existing algorithms, especially when identifying severe bottleneck congestion.

Same-Day Mode Choice Modeling with Household Vehicle Usage Simulation in Developing Countries

This paper presents a model of same-day mode choice at the household level for developing countries. A rule-based algorithm combining classical random utility maximization theory within a microsimulation framework is used. Modeling of private vehicle usage (including vehicle allocation and sharing use in household) is an essential component of this model because vehicle deficiency is common in developing countries. This model consists of four steps: (a) the allocation of private vehicles (car, motorcycle, and bicycle) in a household, (b) the mode choice of private vehicle users specified in the first step, (c) vehicle sharing in a household, and (d) the mode choice of individuals who do not use private vehicles. The adaptability of the model was improved by simulations on car, motorcycle, and bicycle usage. Discrepancies in the mode choice behavior of household members with and without the use of private vehicles are captured in this paper through different modeling methods. The rule-based algorithm, binary logit model, multinomial logit model, and mixed logit model were applied together in this four-step model. Travel diary survey data from 2007 from Bengbu, China, were used as an example for the validation test of this model. The results demonstrate that this model can accurately predict the mode choice of all household members in an internally self-consistent and theoretically credible manner for a midsize city in China. The proposed model is highly conducive to travel demand forecasting and transportation policy making.