UF Faculty & Student Research Participation at the 95th Annual Meeting of TRB

University of Florida was well represented at the 95th Annual Meeting of the Transportation Research Board (TRB). Nearly forty faculty and students were represented in almost thirty papers and presentations. Please see below for the full list, including abstracts.

Analysis of Advanced Management of Curbside Parking

(Presentation #16-1107)

Zhibin Chen, University of Florida
Zhengtian Xu, University of Florida
Mahmood Zangui, University of Florida
Yafeng Yin, University of Florida

This paper analyzes drivers’ search behavior for curbside parking on a one-way street and develops analytical models to compare the impacts of parking information and reservation services. An agent-based simulation experiment is conducted to verify the results drawn from the analytical models. Results show that the reservation service can improve the performance of the parking system, while providing information may not. The former is proved to be a better policy than the latter in the particular setting of parking competition and search behaviors considered in this paper.

Analysis of Structural Behavior of Precast Prestressed Concrete Pavement Using Finite Element Analysis

(Presentation #16-1892)

Kukjoo Kim, University of Florida
Mang Tia, University of Florida
James Greene, Florida Department of Transportation

A full-scale Precast Prestressed Concrete Pavement (PPCP) system was constructed and evaluated under actual traffic load condition to develop the design guideline under Florida condition. This test section shows good load transfer efficiency (LTE) and riding quality. However, there was a lack of information about its structural response and potential performance. A 3-D finite element model was developed for stress analysis of PPCP under critical loading conditions. The developed 3-D model was calibrated using FWD deflection data and used to perform a parametric analysis to determine the effects of critical loading location, concrete modulus, coefficient of thermal expansion of concrete, loss of prestress force, and subgrade stiffness under typical Florida conditions. Results of the parametric study indicate that the maximum stresses in the concrete were observed to increase significantly as the concrete modulus and coefficient thermal expansion increase. Due to the increase in flexural strength associated with the increase in elastic modulus of the concrete, an increase in elastic modulus of the concrete results in a decrease in the computed stress-to-strength ratio under critical temperature-load conditions. The PPCP system evaluated appear to have a good predicted pavement performance with computed stress-to-strength ratio of less than 0.5 with up to additional loss of 20 % of prestress force in the longitudinal and transverse directions. Variations in the base and subbase properties were found to have minimal effect on the maximum induced stresses in concrete, which indicate that the PPCP system is appropriate for a wide variety of subbase and subgrade conditions.

Cracking Performance of Asphalt Mixtures with Alternative Polymer Modified Asphalt Binders Based on Binder Fracture Energy Density

(Presentation #16-1888)

Yu Yan, University of Florida
Sanghyun Chun, University of Florida
Reynaldo Roque, University of Florida
Sungho Kim, Virginia Department of Transportation

This study evaluated the fracture properties of unmodified, styrene-butadiene-styrene (SBS) polymer-modified asphalt (PMA) binder and alternative PMA binders, as well as the fracture properties of hot mix asphalt (HMA) mixtures with the same asphalt binders. Both multiple stress creep recovery (MSCR) and binder fracture energy (BFE) tests identified the unmodified binder, and the alternative PMA binders were found to have comparable or even higher binder fracture energy density (FED) than the PG 76-22 SBS PMA binder. Also, polymer modification, including SBS and alternative polymer, significantly reduced the rate of damage accumulation of asphalt mixtures compared to the one with unmodified binder and resulted in higher energy ratio (ER) which indicates better mixture cracking performance. In particular, mixtures with alternative PMA binders showed relatively higher mixture FED and ER than the one with PG 76-22 SBS PMA binder which validates the feasibility of substituting PG 76-22 SBS PMA with those alternative PMA binders with no compromise on cracking performance of asphalt mixtures. Also, a good correlation was observed between binder FED and mixture FED and consequently, ER values clearly indicated that BFE test can be used as a simple but effective tool to properly characterize and optimize the fracture properties of asphalt mixture for consistently enhanced cracking performance.

Crash Prediction Method for Freeway Segments with High Occupancy Vehicle (HOV) Lanes

(Presentation #16-6333**)

Sivaramakrishnan Srinivasan, University of Florida
Phillip Haas, University of Florida
Priyanka Alluri, Florida International University
Albert Gan, Florida International University
James Bonneson, Kittelson & Associates, Inc.

This study developed methods for estimating the expected crash frequency on urban freeway segments with HOV lanes. The safety impacts of the type of separation between the managed lanes and general purpose lanes were examined. Separate models were estimated for fatal and injury (FI) crashes and all crashes using five years’ of data from California, Washington and Florida. All these facilities have one HOV in each direction (included in the count of total number of lanes). Models are presented for six-, eight-, ten-, and twelve- lane facilities. The effect of separation type on crash rates is found to be statistically significant only in the models for ten-lane facilities. All the estimated models have been implemented in a spreadsheet program which will enable analysts to apply these equations for crash prediction. Overall, this study provides procedures that will help Florida Department of Transportation (FDOT) consider safety in decisions about planning and designing freeways with HOV or HOT lanes.

Development of an 1100C Crushable Nose Surrogate Vehicle for Low Speed Impact Testing of Breakaway Hardware

(Presentation #16-4270)

Matías Groetaers, Walter P Moore
Gary Consolazio, University of Florida
David Wagner, Florida Department of Transportation

In the development of breakaway roadside safety hardware, surrogate vehicles offer an efficient, cost-effective means of testing new breakaway design concepts and evaluating system performance in accordance with applicable standards. For purposes of conducting low speed impact testing, a crushable nose impactor, when combined with a gravity pendulum, can constitute a suitable surrogate alternative to a production vehicle. This paper describes the development of a crushable nose impactor that can serve as a surrogate version of the 1100C passenger car described in AASHTO MASH. Design of the new crushable nose was based on published force-deformation data collected from crash testing of a Kia Rio production passenger car. The new crushable nose differs from other designs in that it is intended specifically for use in low speed impact testing of breakaway hardware; it uses tapered (i.e., trapezoidal) aluminum honeycomb cartridges which permits improved approximation of the production vehicle force-deformation curve; and the cartridge design process incorporates insights gained through high-deformation compression testing of aluminum honeycomb materials. Analysis techniques used to iteratively design the new crushable nose are documented, as are the final dimensions of the configured cartridges. Results from rigid pole validation impact testing of the crushable nose impactor, conducted using the FDOT gravity pendulum, are presented and are shown to be in good agreement with the production vehicle force-deformation data.

Effect of Average Passing Zone Length on Spanish Two-Lane Highways Traffic Performance

(Presentation #16-0169)

Ana Moreno, Universitat Politecnica de Valencia
Carlos LLorca, Universitat Politecnica de Valencia
Scott Washburn, University of Florida
Jose Bessa, Federal Center for Technological Education of Minas Gerais – CEFET-MG
Alfredo Garcia, Universitat Politecnica de Valencia

The U.S. Highway Capacity Manual (HCM 2010) methodology is used in Spain to evaluate traffic operation and quality of service. Recently, the procedure for two-lane highways was adapted to Spanish conditions. The methodology is largely based on simulations from TWOPAS, which was calibrated and validated with Spanish field data, including the number of passes. However, neither the U.S. nor Spanish procedures consider the possible effect of average passing zone length on traffic performance.

The objective of this research is to evaluate the effect of average passing zone length on two-lane highway operation and incorporate it in the analysis methodology, if significant. The TWOPAS microsimulation program was calibrated and validated. Then, it was applied to different scenarios varying the average passing zone length (250 to 5,000 m), directional traffic flow rate, and directional split. Only scenarios with 50% of no-passing zones over the length of the study segment were analyzed.
Average passing zone length had a profound effect on traffic performance: PTSF could be increased up to 28% between long and short average passing zone lengths (PZL). Very short passing zones (250 m) did not contribute to operational efficiency; and the effect was stabilized for PZL longer than 2,500 m. In fact, increasing PZL from 1,000 to 5,000 m improved PTSF by only about 5% and ATS about 2 km/h.

The HCM 2010 evaluation procedure estimated fairly good ATS and PSTF on balanced flows and passing zones longer than 1,250 m. In other cases, the error between the HCM estimates and the simulation results was up to 22% and similar to scenarios with 100 % of no passing zones. Therefore, some adjustments to ATS and PTSF depending on PZL should be included. The results could be used to better estimate traffic operation on Spanish two-lane highways and should be added to current methodologies.

Effectiveness of Crack Relief Techniques to Mitigate Reflective Cracking in Asphalt Overlaid Concrete Pavement

(Presentation #16-1609)

Sanghyun Chun, University of Florida
Abdenour Nazef, Florida Department of Transportation
Edward Offei, Applied Research Associates, Inc. (ARA)
Bouzid Choubane, Florida Department of Transportation
James Greene, Florida Department of Transportation

This study focused primarily on evaluating the effectiveness of five crack relief treatments to mitigate reflective cracking of asphalt overlaid concrete pavement including 0.5 inch Superpave 9.5 mm Nominal Maximum Aggregate Size (NMAS) structural course (SP-9.5), 1.5 inch SP-12.5, 2.5 inch SP-12.5, 1.0 inch Open-Graded Crack Relief (OGCR), and 0.5 inch Asphalt Rubber Membrane Interlayer (ARMI). Pavement performance was evaluated in terms of deflection, ride quality, rutting and cracking. Also, an evaluation of cracking performance based on the dissipated energy concept using Falling Weight Deflectometer (FWD) load-deflection time history data was presented. The results indicated that 0.5 inch SP-9.5, 1.5 inch SP-12.5, and 2.5 inch SP-12.5 treatments show relatively greater potential for reflective crack mitigation efficiency than the 1 inch OGCR or 0.5 inch ARMI. More reflective transverse cracks corresponding to the Portland Cement Concrete (PCC) base slab joints occurred in areas with thinner Asphalt Concrete (AC) overlay and were most predominant in the 0.5 inch ARMI section. Areas with more transverse cracks had relatively greater dissipated energy indicating a higher potential for pavement deterioration and/or damage rate. Therefore, the dissipated energy approach can be used as a reliable indicator of pavement cracking performance. Thinner AC overlay may reduce the reflective cracking mitigation potential of crack relief layers.

Evaluation of Photometric Stereo and Independent Component Analysis Method for Characterizing Aggregate Polishing Behavior

(Presentation #16-6244)

Abolfazl Ravanshad, Fugro Consultants, Inc.
Filippo Fabbi, University of Parma
Reynaldo Roque, University of Florida

Polishing behavior has long been served as a useful criterion for aggregate pre-evaluation. In the last decade, hybrid application of digital image processing and accelerated polishing techniques have been widely used to provide a quick and reliable tool for characterizing aggregate polishing behavior in laboratory. This application is of interest to many highway agencies, as it can lead to considerable saving in quality assurance costs. Previously, we studied use of Aggregate Image Measurement System (AIMS) and Micro-Deval to evaluate frictional characteristics of a broad range of Florida aggregates. We showed that AIMS has a limitation in texture analysis associated with the effect of color variation. Recently, we developed a novel method with the motivation of finding a practical solution to the issue. The proposed methodology is based on combined application of photometric stereo and independent component analysis, and hence is called PS-ICA. The current study was mainly focused on evaluating PS-ICA method using large number of virgin and polished granitic and limestone aggregates. Laboratory experiment was conducted to compare the AIMS and PS-ICA methods along with an independent experimental roughness evaluation using Scanning Electron Microscopy. It was concluded that the PS-ICA method effectively excluded effects of color variations in texture analysis of aggregates. In addition, the results clearly indicated that Texture Index alone does not provide enough information with respect to the surface height distribution, and additional index may be required to better describe aggregate surface microtexture. It is also important to analyze aggregate texture at multiple magnifications.

Evaluation of Public Perception and Outreach Efforts for Roundabouts in Fairmont, West Virginia

(Presentation #16-5298**)

Mehrdad Shahabi, University of Florida
Jenny Elyard
Avinash Unnikrishnan, Portland State University
David Martinelli, West Virginia University
Andrew Nichols, Marshall University
Colin Frosch, West Virginia University
Rachel James, University of Texas, Austin

In 2010, the Gateway Connector in Fairmont, WV was opened to the public; this stretch of roadway, directly connecting Fairmont, WV to the major local interstate, contained the state’s first two roundabouts. A public outreach campaign was organized by Federal Highway Administration, West Virginia Department of Highways, and City of Fairmont officials. This paper presents the public awareness campaign developed for the first two roundabouts constructed in the state of WV in Fairmont. The public perception of the roundabouts was monitored using a progressive survey. The pre-survey was administered shortly before the opening of the roundabout while the post-survey was distributed approximately a year after the roundabout had been in operation. After a year of use, the reported approval rating regarding the public’s perceived comfort of navigating the roundabout and the safety of drivers, pedestrians, and cyclists increased at least 25% between the two surveys. All of the changes in opinion for the survey questions comparing public acceptance of the roundabout, before and after the Connector’s opening, were found to be statistically significant at the 1% significance level.

Examining the Influence of Built Environment on Traffic Crashes: A Spatial Data Mining Approach

(Presentation #16-6530)

Yiqiang Ouyang, University of Florida
Ilir Bejleri, University of Florida

Data mining – a process for extracting implicit, nontrivial, previously unknown and potentially useful information from data in databases – has gained the attention of transportation safety researchers. However, only a few studies have used the association rule data mining for crash analysis. Moreover, the literature that specifically examines the influence of built environment on crashes by applying spatial data mining techniques is very shallow. This research applies a spatial association rule data mining approach to understand the effects of built environment on traffic crashes. This approach consists of five steps: crash data aggregation, calculation of transportation D variables, extraction of spatial topological and distance relationships, numerical data discretization, and rule mining and visualization. The census block-group is selected as the analysis unit for the study. The apriori rule mining algorithm-based R package is used to mine the rules. The results show that many rules were mined for block groups with a large number of crashes. The resulting rules are a mix of D variables, spatial distance and topological variables. This suggests that an interaction of multiple built environment elements influence the occurrence of crashes. Results show that residential and commercial land use types play an essential role in the occurrence of large number of crashes. These land use types influence crashes through spatial distance or topological variables combined with built environment D variables. The result also shows that the Density, Diversity and Design variables are not as important in crash occurrences as the Distance and Destination variables.

Factors Influencing Analytically Derived Barge Force-Deformation Relationships Used in Structural Design

(Presentation #16-2124**)

George Kantrales, University of Florida
Gary Consolazio, University of Florida

Barge bow force-deformation relationships are frequently embedded in barge impact load calculation procedures that are used in the design of bridges and waterway structures. As such, the nonlinear stiffness described by such relationships relates directly to the inherent conservatism of calculated design forces. Since analytical finite element simulation procedures are often used to characterize barge force-deformation relationships, quantifying the conservatism and sensitivity of such relationships is of direct relevance to structural design. In this paper, high-fidelity finite element simulations of barge bow crushing are used to investigate the sensitivity of force-deformation relationships to five modeling characteristics—barge weld pattern, impactor crush speed, strain-rate material parameters, steel failure strain, and steel yield stress. Results from this investigation indicate that by including strain-rate effects, changing weld patterns, or modifying steel yield stress levels, differences in force-deformation behavior, and therefore crush force, are observed. In contrast, changes in crush speed are found to be only weakly correlated to changes in predicted crush force. Findings from this study provide guidance in terms of making appropriate decisions in numerical modeling efforts, and will help to ensure that appropriate levels of conservatism are inherently embedded within impact load calculation procedures used in structural design.

Identification and Selection of Managed Lane Strategies on Arterials

(Presentation #16-1345)

Zhibin Chen, University of Florida
Yafeng Yin, University of Florida
Mahmood Zangui, University of Florida
Chris Francis, Florida Department of Transportation

Although managed lanes have been widely deployed along limited-access facilities to mitigate congestion or improve mobility, they are seldom deployed in arterials. In this study, we identify the types of managed lanes and their variants that have potential for managing traffic along arterials. Our investigation suggests that high-occupancy vehicle, high-occupancy/toll, express-toll, and bus-only lanes are more suitable for implementation on arterials. Further, details concerning their designs and implementations are discussed along with traffic management schemes such as intersection treatment, segment management, and automatic enforcement. Lastly, a flow chart is provided for determining an appropriate managed lane strategy, including types, designs and methods of implementation, and traffic management schemes, for implementation.

Impact of Hydrated Lime on Cracking Performance of Asphalt Mixtures with Oxidation and Cyclic Pore Pressure

(Presentation #16-3745)

Jian Zou, University of Florida
Reynaldo Roque, University of Florida
George Lopp, University of Florida
Marco Isola, University of Florida
Michael Bekoe, University of Florida

This research was conducted to evaluate impact of hydrated lime (HL) on cracking performance of asphalt mixtures for a broad range of aggregate types including both granites and limestones. A recently developed conditioning procedure, long-term oven aging (LTOA) plus cyclic pore pressure conditioning (CPPC), was employed to effectively simulate effects of long-term field aging on asphalt mixtures. The energy ratio (ER) approach was used to evaluate cracking performance of untreated and lime-treated mixtures, which takes into account mixture properties that are known to be closely related to cracking and was calibrated using cores taken from field-aged pavements throughout Florida. The results showed that introduction of HL appeared to mitigate detrimental effects of oxidative aging and moisture (LTOA plus CPPC) on damage and fracture-related properties of both granite and limestone mixtures. This effect resulted in better cracking performance of these mixtures. Further analysis based on comparison of initial construction cost of conventional pavement structure revealed that use of HL can result in less cost to achieve pavements with the same level of good performance. Furthermore, this research showed that the equivalent ER criterion proposed and implemented in the cost analysis can effectively assess the impact of HL-modification.

Impacts of Shoulder Days on Nonmandatory Activity Duration: Time-Use Analysis Based on 2009 NHTS

(Presentation #16-2834)

Hamidreza Asgari, Florida International University
Xia Jin, Florida International University
Sivaramakrishnan Srinivasan, University of Florida
Fengjiang Hu, Florida International University

This study is intended to explore the temporal variations observed in non-mandatory activity involvement from a time-use analysis perspective. Five non-mandatory activities including “Household maintenance”, “Personal maintenance”, “Social”, “Meal” and “Exercise” are examined in terms of activity duration. More importantly, this study emphasizes on the impacts of shoulder days (Mondays and Fridays) on non-mandatory activity duration as there has been a tendency to ignore the roles of these two days of the week compared to typical mid-weekdays. A structural equations model (SEM) is developed to investigate 1) How non-mandatory activities compete with each other in view of time budget allocation, 2) How activity durations very across different segments (days) of the week, and 3) How observed differences could be correlated with socioeconomic and demographic attributes. Temporal variations are captured through the incorporation of a weekday category variable (Monday, Midweek, and Friday), and the associated interaction effects with various personal and household characteristics. The model results reveal significant interaction effects, confirming the hypothesis that people do show temporal preferences on shoulder days compared to typical mid-weekday for non-mandatory errands. The data applied in this study was obtained from the 2009 National Household Travel Survey (NHTS).

Mechanical, Laser, and Plasma Marking on the Fatigue Properties of Bridge Steel

(Presentation #16-6076**)

Michele Manuel, University of Florida
Michael Kesler, University of Florida
Peter Feldtmann, University of Florida
Edward George, E and S Consulting
Steven Duke, Florida Department of Transportation

Scribe marks on high-strength low-alloy (HSLA) weathering steels used for non-redundant fracture critical members in bridge construction are generally stated as prohibited in most state Departments of Transportation (DOTs) guidelines due concerns surrounding the degradation of mechanical properties. Currently, any automated scribing marks allowed, namely mechanical milling, are either cut from the end of the part, or subsequently welded over, thereby effectively removing it prior to service. All other markings are either manually die-stamped or spray-painted on. The lack of an automated capability to place markings on weathering steels slows production and markings are often accidentally removed during sand blasting or shipping, which causes additional problems for the manufacturer and recipient. There is a need to establish safe, automated methods of scribing fracture critical members such that markings will remain throughout the production process, but will not compromise the integrity over the lifetime of the part. Three automated technique are of interest as they are often already in the manufacturing process line. These include: mechanical milling, plasma scribing and laser scribing. In this study, a microstructural evaluation of the bulk material and the changes which occur around the three markings are investigated. Plasma scribing resulted in the most noticeable surface marking. S-N curves generated through fatigue testing showed no measurable difference in fatigue life between marked and unmarked material.

Methodological Approach to Develop a Comprehensive Transportation Accessibility Index for the Transportation-Disadvantaged

(Presentation #16-6209)

Ilir Bejleri, University of Florida
Soowoong Noh, University of Florida
Zongni Gu, University of Florida

A certain segment of population, such as elderly, disabled and people that don’t own vehicle, referred to as transportation disadvantaged (TD), face mobility challenges due to lack of alternative transportation options. To determine and communicate comprehensive transportation deficiencies to policy makers, the evaluation of transportation accessibility for TD should include all modes such as public transportation, on-demand services, and others. Previous studies in this area have typically been focused on one aspect of transportation accessibility or primarily have illustrated broad geographic gaps at state or county level. This study develops a GIS-based methodology to calculate a comprehensive transportation accessibility index that takes into account public transportation, on-demand services, and or special services. Geographic service areas are created for each category based on network travel distances, bus routes and stations, and a serviceability index by census block groups. Each service layer is normalized using standardized scores and the three layers are merged into one composite layer, which presents a comprehensive transportation accessibility measurement. In addition, the method can determine service gaps – high demand and low supply – by overlaying TD location layers with the composite transportation accessibility index. This method is specifically targeted to assessing accessibility of TD more comprehensibly and at a finer geographic scale. The GIS model presented can be used to identify areas that are not adequately serviced with alternative transportation options for TD, and guide policy makers to make decisions that foster equitable provision of transportation. Alachua County in Florida is selected as test bed for application and evaluation of the methodology presented in this work.

Multimodal Transportation Choices and Health: Exploratory Analysis Using Data Fusion Techniques

(Presentation #16-6320**)

Miguel Lugo, University of Florida
Sivaramakrishnan Srinivasan, University of Florida

In this study, we demonstrated the feasibility of using data fusion in the context of large-scale travel and health surveys, and subsequently used the new comprehensive dataset generated to model the relationship between health and multi-modal (walking, biking, transit, and vehicle usage) long-term (weekly/monthly/yearly) travel choices. Two measures of health are fused from a helth survey on to a travel survey at the disaggregate (individual) level. The probabilistic record linage software Link Plus was used for the data fusion purposes. The methodology was validated using the EH module of the American Time Use Surveys. Subsequently, the algorithm was used to match the health information from the ATUS to the NHTS and the resulting master dataset was used to develop models for multi-modal travel choices and health. The statistical analysis indicates that while incrasing walking and transit use is associated with better health (relative to non users of the mode), those with the highest levels of walking and transit use are also in poor health relative to moderate users of the mode. Similarly, those at the two ends of the driving spectrum (first and fourth quartiles) have higher BMI compared to those in the middle of the spectrum. There were no statistically significant effects of weekly bike trips on health measures. Overall this study is envisioned as a proof of concept of how data fusion techniques may be used to integrate multiple datasets to facilitate a comprehensive study of multi-modal travel choices and health.

Performance Measures on Two-Lane Highways: Survey of Practice

(Presentation #16-5259**)

Ahmed Al-Kaisy, Montana State University
Amirhossein Jafari, Montana State University
Scott Washburn, University of Florida
Tapio Luttinen, Aalto University
Richard Dowling, Kittelson & Associates, Inc. (KAI)

Two-lane highways constitute a significant component of the highway system and serve an essential function for the movement of people and goods. The Highway Capacity Manual is primarily used for operational analysis on two lane highways in the U.S., Canada and a few other countries outside North America. As part of a project sponsored by the National Cooperative Highway Research Program for improving the two-lane highway operational analysis methodology, a practice survey was conducted to gather information on agency experiences and preferences in regards to performance measurement on two-lane highways. The survey was sent to all state Departments of Transportation (DOTs) in the United States and Canada. A total of 35 usable responses were received, representing 24 states and four Canadian provinces, with a few agencies sending multiple responses. This paper summarizes the results of the agency survey and presents the most important findings. Results suggest that the top three criteria for good performance measures on two-lane highways are: sensitivity to traffic conditions, sensitivity to road conditions, and relevance to road user perception. Further, agencies identified average travel speed as the most relevant traffic flow aspect to two-lane highway operations. Other performance measures that were found meritorious were volume-to-capacity ratio and flow rate, for class I and class II highways, respectively, versus average travel speed, volume-to-capacity ratio, and percent-time-spent-following for class III highways.

Predicting Subway Volume Using Local Linear Kernel Regression

(Presentation #16-0216)

Yu-Chen Yang, National Sun Yat-Sen University
Chao Ding, University of Hong Kong
Jie Cao, National Sun Yat-Sen University
Yong Jin, University of Florida

Our study is entrusted by the Kaohsiung City Government in Taiwan and the Kaohsiung Rapid Transit Corporation. We attempt to devise a more effective methodology to forecast the passenger volume of the subway system in the city of Kaohsiung. A local linear kernel model is proposed to incorporate different weights for each realized observations. It enables us to capture richer information and improve rate of accuracy. We compare different methodologies, for example, ARIMA, Best in-sample fit ARIMA, linear model, and their rolling versions with our proposed local linear kernel regression model by examining the in-sample and out-of-sample performances. Our results indicate that the proposed rolling local linear kernel regression model performs the best in forecasting the passenger volume in terms of smaller prediction errors in a wide range of measurements.

Preliminary Cost Estimation of Resurfacing Projects Using Penalized Regression: The Lasso Method

(Presentation #16-0796)

Yuanxin Zhang, University of Florida
R. Edward Minchin, University of Florida
Lourdes R. Ptschelinzew, University of Florida

The quality of preliminary cost estimates has been widely recognized as a critical factor for both State Highway Agencies (SHAs) and highway contractors in decision making at the early stages of transportation projects. The conventional detailed estimating approach is not suitable for early cost estimates not only because it is time-consuming to create an accurate estimate, but also because it heavily depends on the information available-information that is greatly lacking at the very early stages. Parametric estimating methods are effective alternatives that can offset the shortcomings associated with detailed estimating. The most commonly used parametric estimating methods identified in literature are ordinary least squares (OLS) regression and artificial neural networks. However, these approaches have their limitations. This paper proposes a novel approach – Least Absolute Shrinkage and Selection Operator (LASSO) – that is significantly better than the OLS method in many aspects, including automatic feature selection, the ability of dealing with highly correlated data, easier interpretability, and numerical stability of the model predictions. This study applied the LASSO approach to resurfacing project data gathered from the Florida Department of Transportation. The data were separated into two groups: one for training, and the other for validation purposes. Both LASSO and OLS methods were used to build multivariate regression models and then the models’ performances were evaluated based on the Root Mean Square Error (RMSE). The results showed that the LASSO regression model outperformed the OLS regression model. The proposed procedure can also be used to estimate the cost of other types of project at the early phases.

Pricing Strategies for an E-hailing Platform in Taxi Service

(Presentation #16-4914)

Xiaolei Wang, Shanghai Jiao Tong University
Fang He, University of Florida
Huaizhu Gao, Cornell University

Taxi market has become a more complex system with the rapid development of E-hailing, which supplements the conventional roadside hailing, for the provision and usage of taxi service. Pricing schemes of an E-hailing platform for both taxi service providers and users are crucial for platform profitability as well as the overall performance of the taxi market. In this paper the authors conduct a first quantitative investigation into fundamental questions of 1) how an E-hailing platform’s pricing strategy influences the taxi market equilibrium; 2) what the best pricing strategies would be; and 3) whether and under what circumstances can we find Pareto-improving (PI) pricing schemes that improve one mode with sacrificing the other. Based on the complex relationships among demand, supply and efficiency variables of each mode, we derive the partial derivatives of the taxi market equilibrium with respect to the platform’s price perturbations. Such price perturbations are then translated into market variability of taxi service, changes in the profitability of the e-hailing platform, and implications for social welfare, which further enables us to examine optimal pricing strategies from both public and private E-hailing platforms’ perspective. Based on the derivative information, we also theoretically exclude the existence of PI perturbations to the platform’s pricing strategies when both modes exhibit the same type of returns to scale, and establish necessary and sufficient conditions for the existence of PI strategies in the other cases.

Probabilistic Analysis to Quantify the Effect of Use of a Material Transfer Device on Smoothness Pay Incentive

(Presentation #16-5321)

Ohhoon Kwon, University of Florida
Hyung Suk Lee, Applied Research Associates, Inc. (ARA)
James Greene, Florida Department of Transportation
Bouzid Choubane, Florida Department of Transportation
Richard Hewitt, Florida Department of Transportation

The benefit of a Material Transfer Device (MTD) during asphalt paving is well known for providing smoother pavement while eliminating thermal and material segregation. However, the additional operating cost may prevent contractors from investing in the use of an MTD. This study evaluates the effect of MTDs on the smoothness of the finished friction course layer and predicts the incentives/disincentives a contractor may receive using an MTD. Two empirical models were developed using simple linear regression analysis to predict final friction course smoothness based on the underlying structural layer smoothness and MTD use. The regression model results utilize Monte Carlo simulation in conjunction with residual analysis techniques to predict the incentives/disincentives that a contractor may receive per FDOT’s new Developmental Specification for smoothness in a probabilistic fashion. Finally, expected incentives were predicted as a simulation result. Two validation projects with a normalized length of 10 miles showed additional expected incentives of $9,700 and $10,000 when an MTD was used compared to the expected incentives when an MTD was not used. It is noted that the additional incentives are approximately 75% more than the typical operating cost of an MTD for a 10 lane mile friction course construction in Florida.

Roundabouts as a Form of Access Management

(Presentation #16-6236)

Dimitra Michalaka, The Citadel
Ruoying Xu, University of California, Berkeley
Joseph Page, The Citadel
Ruth Steiner, University of Florida
Scott Washburn, University of Florida
Lily Elefteriadou, University of Florida

Transportation engineers and planners are becoming more interested in using roundabouts to address capacity and traffic operations, access management and safety concerns in the transportation system. Roundabouts are preferred as a form of intersection by some communities because they allow for continuous flow of traffic and they can be less expensive to maintain and operate. However, the operational characteristics in the functional area are less well understood than other types of intersections. Therefore, while roundabouts are being used in a variety of contexts, existing research does not provide detailed guidance on how to evaluate the use of roundabouts as a form of access management. This paper provides an operational analysis of selected roundabouts in Florida and South Carolina. The operational analysis identifies four areas of concern: corner clearance, including stopping site distance (SSD) and intersection sight distance (ISD); the flow of traffic and the queuing of vehicles in the functional area near roundabouts, including driveway and intersection spacing, and the use of medians; access to major activity centers; and safety of vulnerable road users, especially bicyclists and pedestrians.

Safety-Related Analysis of the 3D Driver Body Posture Using Naturalistic Data

(Presentation #16-2239)

Alexandra Kondyli, University of Kansas
Angelos Barmpoutis, University of Florida
Virginia Sisiopiku, University of Alabama, Birmingham

Although significant advances have been done with respect to vehicle technology and roadway construction, driver behavior remains the number one contributing factor of traffic crashes worldwide. Studies show that one of the major causes of crashes is driver inattention. Driver inattention may occur when drivers are involved with secondary activities (e.g., texting, talking on the phone, or eating), and when they fail to follow the cues of the surrounding environment while driving. The latter is particularly important when drivers are negotiating maneuvers and are required to interact with other vehicles as in the case of changing lanes or merging onto the freeway. The main objective of this research is to investigate the relationship between driver behavior and safety, by looking at the actual body movements and posture, as well as the eye fixation of the drivers when they are performing lane changing and merging maneuvers under different traffic conditions. To accomplish this objective, a total of 35 drivers were recruited to participate in an instrumented vehicle field study, where each participant drove for approximately two hours along a pre-selected route. Participants’ 3D body posture was recorded with the use of a low-cost infrared depth sensor (Microsoft Kinect). Participants’ eye gaze throughout the entire data collection effort was recorded with the help of eye-tracking equipment. The vehicle was equipped with two cameras that faced the front and the rear, and therefore, information about the traffic conditions during the data collection period was obtained. A rich dataset of driver behavior was developed and analyzed as part of this research. The analysis findings relate the 3D sequence of driver motion and posture with the actual eye and head movement of drivers. This paper presents the research approach, summarizes findings, and provides recommendations accordingly. The findings are expected to assist in establishing monitoring guidelines for advanced driver assistance systems that take into account the driver’s body position and movements, rather than considering solely the vehicle position relative to other vehicles on the road. The results can also assist in developing appropriate alert mechanisms for increasing driver alertness.

Socially Optimal Replacement of Conventional Vehicles with Electric Vehicles: Results for U.S. Household Fleet

(Presentation #16-1389)

Eleftheria Kontou, University of Florida
Yafeng Yin, University of Florida
Zhenhong Lin, Oak Ridge National Laboratory
Fang He, University of Florida

This study presents a dynamic programming framework for minimizing the 15-year societal cost of replacing household, conventional, internal combustion engine vehicles with battery electric ones. The program determines the optimal replacement rate of conventional with electric vehicles, the all-electric driving range that should be market available and the optimal distance between chargers that should be deployed by the government. The framework is applied to datasets that represent the US households, the automobile and the energy market. Results indicate that the optimal percentage of conventional vehicles replaced with electric ones at the end of the 15-year horizon is 30.45%. Socially optimal all-electric driving range is 156-mile, with chargers placed every 184 miles. The decision variables are sensitive to the growth rate of battery pack cost, gasoline, home, and public charger costs. EV diffusion rates increase when gasoline costs increase, electricity cost decreases, and battery pack costs decrease.

Spatiotemporal Pattern Analysis of Taxi Trips in New York City

(Presentation #16-6511)

Hartwig Hochmair, University of Florida

A growing number extensive datasets provides transportation planners with the necessary means to analyse urban travel patterns and gain insight into urban dynamics. This paper explores the spatial and temporal variation of taxi trips in New York City (NYC) over a two month period by analyzing 29 million trip records from a freely available dataset. The study examines the change in trip demand as well as the variation of travel speed at different times of the day for different districts within the study area. Comparison of trip frequencies between weekday and weekend in each district reveals similarities and differences in functional drivers of taxi trip demand and allows to identify different functional areas within the city. A negative binomial regression model that predicts the number of taxi trips per district from the number of rapid transit stations, the location of airports, and the subway ridership provides insight into the correlation between taxi trip demand and public transit infrastructure and ridership. Combining subway ridership and taxi count numbers for each district in a modemix variable allows, using Local Indicators of Spatial Association statistics in a Geographic Information System (GIS), to identify districts that exhibit an increased inclination towards taxi use and are poorly served by public transit. The presented approach could be used as a decision support tool for deciding where investments in rapid transit infrastructure and service would be particularly beneficial for increasing transit more share.

Tradable Credit Scheme to Control Bottleneck Queue Length

(Presentation #16-5687)

Nima Shirmohammadi, University of Florida
Yafeng Yin, University of Florida

This paper proposes to use tradable credits to control the maximum queue length behind a bottleneck. Given a queue length threshold specified by the traffic authority, we design a tradable credit scheme to maintain the queue length of the bottleneck to be less than the threshold for homogeneous travelers. The scheme is compared against a multi-step tolling scheme under demand uncertainty. We then study the capability of the tradable credit scheme in the presence of two types of user heterogeneity: proportional heterogeneity and -heterogeneity. It is found that the proposed scheme can successfully maintain the queue length under proportional heterogeneity, but not necessarily do so under -heterogeneity. However, if the distribution of -heterogeneity is known in advance, the scheme can be modified to ensure the success in controlling queue length.

Variable versus Fixed Congestion Pricing under Day-to-day Traffic Dynamics

(Presentation #16-1411)

Zhengtian Xu, University of Florida
Yafeng Yin, University of Florida
Hai Yang, Hong Kong University of Science and Technology
Feng Xiao, Southwest Jiaotong University

This paper compares the effectiveness of fixed and variable congestion pricing on evolving network flows to a target flow distribution under day-to-day traffic dynamics. Fixed pricing charges constant tolls while variable pricing updates tolls every day depending on recent days’ traffic conditions. With a linear approximation of the dynamical equations, the asymptotic analysis in this paper proves that variable pricing does not necessarily perform better than fixed pricing and subsequently provides conditions when this situation can arise. Numerical experiments are presented to demonstrate the comparison.