2018 Transportation Research Board Annual Meeting

Please join us!

TransInfo will again be hosting a reception at TRB this year. Please join us on Monday, January 8th from 7:30 - 9:30 PM in the West Overlook Room at the Walter E. Washington Convention Center. We hope to see you there for food, drinks and networking! 

ISTL Faculty and Students: Posters, Papers and Presentations

Check out all of the sessions University at Buffalo faculty and students are participating in below. 

Monday, January 8th, 2018

Poster: Innovations in Adaptive Signal Control

Monday 8:00 AM- 9:45 AM

Online Traffic Signal Coordination with a Game Theoretic Approach

Xuan Han, University at Buffalo

Jun Zhuang, University at Buffalo

Qing He, University at Buffalo

Abstract: The occurrence of Connected and Automated Vehicles (CAV) technologies brings an opportunity to develop an online data-driven signal coordination method which cooperatively optimizes the traffic performance of a group intersections. This research proposes a game theoretic approach to epsilon-equilibrium (or near-Nash equilibrium) to achieve online traffic signal coordination. Each intersection of the traffic network is just like a player in a game. Different intersections pursue their own benefit maximization by changing their offsets. This paper examines the effect, applicability, and efficiency of the game theoretic approach in signal coordination. The game theoretical approach is proven to outperform the system optimum on vehicle delay at intersection level regarding delay equity. The variances of vehicle delay among different intersections are significantly decreased by the proposed game theoretic algorithm. Thus no intersection needs to sacrifice its own delay performance to achieve system optimum, and traffic delay has been widely distributed among intersections. In addition, this research also compares the network delay performances between CAVs and Human-driven Vehicle (HDV). A simulation platform is built to evaluate the proposed algorithms and models. The results also show that the CAVs can generate much less delay compared to HDVs.

Poster | Spotlight Theme | Practice Ready Paper: Advances in Urban Freight Transportation Research and Practices

Monday 8:00 AM- 9:45 AM

Using Local Stores for Same Day Delivery

Ming Ni, University at Buffalo

Qing He, University at Buffalo

Jose Walteros, University at Buffalo

Xuan Liu, IBM, Thomas J. Watson Research Center

Arun Hampapur, IBM, Thomas J. Watson Research Center

Abstract: This study investigates the same-day delivery planning with store fulfillment problem (SDD-SFP) that identifies a seasonal order fulfillment plan for delivering local online orders from nearby retailing stores and minimizes the overall planning and operational costs. It aims to develop optimization models and solution algorithms about store location selection, fleet-sizing for transportation, and inventory planning. In order to solve large-scale SDD-SFP with real-world datasets, this paper develops accelerated Benders decomposition methods that integrate outer search tree and local branching based on mixed-integer programming, and optimization-based algorithms for initial lifting constraints

Poster | AICP Certification | Practice Ready Paper: Innovations in Urban Data and Mobility Modeling

Monday 10:15 AM- 12:00 PM

Understanding the Heterogeneity of Human Mobility Patterns: User Characteristics and Modal Preferences

Laiyun Wu

Samiul Hasan, University of Central Florida

Jee Eun Kang, University at Buffalo

Younshik Chung, Yeungnam University

Abstract: Characterizing individual mobility is critical to understand urban dynamics and develop high-resolution mobility models. Previously, large-scale trajectory datasets have been used to characterize universal human mobility patterns. However, due to the limit of the underlying datasets, these studies could not investigate how mobility patterns change over user characteristics. In this paper, we analyze a large-scale Automatic Fare Collection (AFC) dataset of the main transit system of Seoul, South Korea and investigate how mobility patterns change over user characteristics and modal preferences. We identify users' commuting locations and estimate the statistical distributions required to characterize their spatio-temporal mobility patterns. Our findings show the heterogeneity of mobility patterns across demographic user groups. This result will significantly impact future mobility models based on trajectory datasets.

Lectern: Research Advances in Statistical and Econometric Methods

Monday 10:15 AM- 12:00 PM

Panagiotis Anastasopoulos, University at Buffalo, presiding

Lectern: Toward Sustainable and Resilient Transportation Networks

Monday 1:30 PM- 3:15 PM

A Spectral Risk Measure in Hazardous Materials Transportation

Liu Su, University of South Florida

Longsheng Sun, University at Buffalo

Mark Karwan, University at Buffalo

Changhyun Kwon, University of South Florida

Abstract: Due to catastrophic consequences of accidents by hazardous materials (hazmat) transportation, a risk-averse approach for routing is needed.Due to catastrophic consequences of accidents by hazardous materials (hazmat) transportation, a risk-averse approach for routing is needed.In this paper, we consider spectral risk measures, which are coherent and more general than existing approaches such as conditional value-at-risk.In spectral risk measures, one may define the spectrum function to reflect the decision maker's risk preference precisely. We first consider a special class of spectral risk measures, for which the spectrum function may be represented as a weighted sum of step functions.We develop a mixed integer programming model in hazmat routing to minimize these special spectral risk measures and propose an efficient search algorithm to solve the problem.We also consider general classes of spectral risk measures, for which we propose two computational approaches.  We illustrate the usage of spectral risk measures and the proposed computational approaches via a case study in the real road network of Ravenna, Italy.

Lectern: Monitoring and Predicting Surface Conditions Using Road Weather Information Systems and Mobile Sensors

Monday 1:30 PM- 3:15 PM

The Optimal Location of Road Weather Information System in New York State

Julie Fetzer

Hernan Caceres Venegas, Universidad Catolica del Norte

Qing He, University at Buffalo

Rajan Batta, University at Buffalo

Abstract: Inclement weather is a threat to the safety of transportation systems as well as the efficiency of their operation. A road weather information system (RWIS) is a network of environmental sensor stations (ESS) that collect a range of real-time data about weather and pavement conditions. These systems can support highway officials and civilians in making more informed transportation safety decisions, particularly in times of adverse weather, by giving them more accurate and localized weather information. This enables the proper maintenance activities to be executed and safety to be restored while using minimum resources. However, because of the range of network characteristics and geographical factors affecting the implementation of ESS, no widely adopted guidelines exist that outline where to implement ESS in a network beyond taking into the physical criteria of an appropriate site, although several methods have been suggested. This paper aims to take a practical approach to solving the location problem of RWIS by proposing a unified multi-objective optimization methodology that takes into account vehicular accident data, vehicle miles traveled, area coverage, access to power and maintenance, and existing ESS. This study produces an exact solution method that produces a Pareto set of multiple efficient solutions. The proposed methodology is applied to a real world case study focused on the deployment of additional ESS in the existing RWIS network across New York State. Further, a sensitivity analysis is conducted to examine the effects of different parameters and a non-preference solution is proposed.

Poster: Artificial Intelligence and Machine Learning Tools for Estimation, Detection, and Prediction Applications in Transportation

Monday 1:30 PM- 3:15 PM

Detecting Traffic Accidents from Social Media Data with Deep Learning

Zhenhua Zhang, State University of New York (SUNY)

Qing He, University at Buffalo

Jing Gao

Ming Ni, University at Buffalo

Abstract: This paper employs deep learning in detecting the traffic accident from social media data. First, we thoroughly investigate the 1-year over 3 million tweet contents related to traffic accidents in two metropolitan areas: Northern Virginia and New York City. Our results show that paired tokens can capture the association rules inherent in the accident-related tweets and further increase the accuracy of the traffic accident detection. Second, two deep learning methods: Deep Belief Network (DBN) and Long Short-Term Memory (LSTM) are investigated and implemented in extracted tokens. Results show that DBN can obtain an overall accuracy of 85% with about 24 individual token features and 20 paired token features. The classification results from DBN outperform those of support vector machine (SVM) and supervised Latent Dirichlet allocation (sLDA). Finally, based on the time and location comparison between accident-related tweets with the traffic accident record on two freeway segments: I66 and I395, it is found that half the accidents detected by Twitter can be found in the accident record.

 

Tuesday, January 9th, 2018

Artificial Intelligence and Advanced Computing Applications Committee

Tuesday 8:00 AM- 12:00 PM 

Sherif Ishak, University of Alabama, Huntsville, presiding

Adel Sadek, University at Buffalo, presiding

Wednesday, January 10th, 2018

Poster | Practice Ready Paper: Social Networks and Group Behavior

Wednesday 8:00 AM- 9:45 AM 

Travel Behavior Classification: An Approach with Social Network and Deep Learning 

Yu Cui, University at Buffalo

Qing He, University at Buffalo

Alireza Khani, University of Minnesota, Twin Cities

Abstract: Uncovering human travel behavior is crucial for not only travel demand analysis but also ridesharing opportunities. To group similar travelers, this paper develops a deep learning based approach to classify travelers’ behaviors given their trip characteristics, including time of day and day of week for trips, travel modes, previous trip purposes, personal demographics, and nearby place categories of trip ends. This study first examines the dataset of California Household Travel Survey (CHTS) between the year of 2012 and 2013. After preprocessing and exploring the raw data, we construct an activity matrix for each participant. The Jaccard similarity coefficient is employed to calculate matrix similarities between each pair of individuals. Moreover, given matrix similarity measures, we construct a community social network for all participants. We further implement a community detection algorithm to cluster travelers with similar travel behavior into the same groups. There are five clusters detected: non-working people with more shopping activities, non-working people with more recreation activities, normal commute working people, shorter working duration people, later working time people, and individuals needing to attend school. We further build an image of activity map from each participant’s activity matrix. Finally, a deep learning approach with convolutional neural network is employed to classify travelers into corresponding groups according to their activity maps. The accuracy of classification reaches up to 97%. The proposed approach offers a new perspective for travel behavior analysis and traveler classification.

Poster | Practice Ready Paper: Activity Patterns and Activity Scheduling

Wednesday 8:00 AM- 9:45 AM 

Multi-day Activity-Travel Pattern Sampling Based on Single-Day Data 

Anpeng Zhang, University at Buffalo

Jee Eun Kang, University at Buffalo

Kay Axhausen, ETHZ - Swiss Federal Institute of Technology

Changhyun Kwon, University of South Florida

Abstract: Although it is important to consider multi-day activities in transportation planning, multi-day activity-travel data are expensive to acquire. In this study, we propose to generate multi-day activity-travel data through sampling from readily available single-day household travel survey data with considerations of day-to-day intrapersonal variability. One of the key observations we make is that the distribution of interpersonal variability in single-day travel activity datasets is similar to the distribution of intrapersonal variability in multi-day datasets. Thus, interpersonal variability observed in cross-sectional single-day data of a large population can be used to generate the day-to-day intrapersonal variability. The proposed sampling method is based on activity-travel pattern type clustering, travel distance and variability distribution to extract such information from single-day data. Validation and stability tests of the proposed sampling methods are presented.

Poster: The Search for a Better Way, Part 2: Exploring and Refining Methods in Highway Safety (Part 1, Session 523)

Wednesday 10:15 AM- 12:00 PM 

Analysis of accident injury-severities using a time-variant correlated random parameters ordered probit approach 

Grigorios Fountas, University at Buffalo

Panagiotis Anastasopoulos, University at Buffalo

Mohamed Abdel-Aty, University of Central Florida

Abstract: This paper employs a correlated random parameters ordered probit modeling framework to explore time-variant and time-invariant factors affecting injury-severity outcomes in single-vehicle accidents.  The proposed approach extends traditional random parameters modeling, by accounting for possible correlations among the random parameters.  On the basis of an unrestricted covariance matrix for the random parameters, the proposed framework can capture the combined effect of the unobserved factors – which are captured by the random parameters – on the injury-severity mechanism.  The empirical analysis is based on traditional roadway-, traffic- and crash-specific information, and detailed weather and pavement surface disaggregate data, collected in the State of Washington, between 2011 and 2013.  The results show that accident injury-severity outcomes are affected by a number of time-variant (ice thickness or water depth on pavement surface, sub-surface temperature) and time-invariant (roadway geometrics, and    vehicle-, driver-, and collision-specific characteristics) factors, several of which result in statistically significant parameters – thus they have mixed effects on the injury-severity generation mechanism.  The findings also present statistically significant correlation effects among the random parameters, which substantiates the appropriateness of the approach.  The comparative assessment between the employed approach and its lower-order counterparts (i.e., fixed parameters, and uncorrelated random parameters ordered probit modeling approaches) shows that accounting for the unobserved heterogeneity interactions results not only in superior statistical performance (in terms of model’s fit, and explanatory and prediction performance) but also in less biased and more consistent parameter estimates.

 

Poster | Practice Ready Paper: Transportation Network Modeling

Wednesday 2:30 PM- 4:00 PM

A Probabilistic Trip Chaining Algorithm for Transit Origin-Destination Matrix Estimation using Automated Data 

Pramesh Kumar, University of Minnesota, Twin Cities

Alireza Khani, University of Minnesota, Twin Cities

Qing He, University at Buffalo

Abstract: Development of an origin-destination demand matrix is crucial for transit planning. The development process is eased with automated data collection, making it possible to mine boarding and alighting patterns on an individual basis. This research proposes a probabilistic method for trip chaining which uses Automatic Fare Collection (AFC) and General Feed Transit Specification (GTFS) data. The proposed method avoids problems resulting from errors in AFC transaction locations or selection of incorrect subroutes from GTFS data which may result in incorrect inference of the origin or destination. The method has been applied to the Twin Cities AFC data as a case study. The transit system is an open system where passengers tap smart cards once while boarding or sometimes alighting (on outbound pay-exit buses). Based on the consecutive tags of the passenger, an algorithm with different capabilities is developed for different cases based on the pay-exit property. The method is compared with a baseline algorithm which shows improvements in the quantity and quality of inferred trips. Finally, the inferred origin-destination demand matrices as well as route ridership are presented visually for planning purposes.

Poster | Practice Ready Paper: Transportation Network Modeling

Wednesday 2:30 PM- 4:00 PM

A Spectral Risk Measure in Hazardous Materials Transportation 

Liu Su, University of South Florida

Longsheng Sun, University at Buffalo

Mark Karwan, University at Buffalo

Changhyun Kwon, University of South Florida