Assessment of the Need for System Warrants in Addition to Local Warrants for Ramp Metering Installation
Homa Fartash, Florida International UniversityShow Abstract
Mohammed Hadi, Florida International University
Yan Xiao, Florida International University
Warrants for ramp metering installation have been developed by a number of states around the nation. These warrants are generally simple and are based on the traffic, geometry, and safety conditions in the immediate vicinity of each ramp (local conditions). However, advanced applications of ramp metering utilize system-based metering algorithms that involve metering a number of on-ramps to address system bottleneck locations. These algorithms have been proven to perform better compared to local ramp metering algorithms. This has created a disconnection between existing agency metering warrants to install the meters and the subsequent management and operations of the ramp meters.
This study focuses on assessing the need for system warrants, in addition to local warrants for ramp metering installation to prevent traffic breakdown at bottleneck locations. A linear programming formulation combined with the consideration of the stochasticity of bottleneck capacity was used to select the ramps to be metered based on the system bottleneck. The study found that the selection of additional ramps for metering based on system bottlenecks, in addition to those justified by local warrants, can delay the breakdown at the system bottleneck location and improve the performance of the freeway mainline. Another important benefit of selecting ramps for metering based on system operations is that it distributes the on-ramp delays due to metering among more ramps. This leads to reducing the experienced delays on the ramps selected utilizing the existing warrants that are based on local conditions.
Large-Scale Traffic Incident Duration Analysis: Role of Multiagency Response and On-Scene Times
Xiaobing Li, University of AlabamaShow Abstract
Asad Khattak, University of Tennessee
Behram Wali, Urban Design 4 Health, Inc.
Traffic incidents often known as non-recurring events impose enormous economic and social costs. Compared to short duration incidents, large-scale incidents can substantially disrupt traffic flows by blocking lanes on highways for long periods of time. A careful examination of large-scale incidents and associated factors can assist with actionable large-scale incident management strategies. For such an analysis, a unique and comprehensive 5-year incident database on East Tennessee roadways was assembled to conduct in-depth investigation of large-scale incidents, especially focusing on operational responses (e.g., response and stay times) by various agencies. Incidents longer than 120 minutes and block at least one lane are considered large-scale, giving 890 incidents, which are about 0.69% of all reported incidents in the database. Rigorous fixed- and random-parameter hazard-based duration models are estimated to account for the possibility of unobserved heterogeneity in large-scale incidents. The modeling results reveal significant heterogeneity in associations between operational responses and large-scale incident durations. A 30-minute increase in response time for first, second, and third (or more) highway response units translates to 2.8, 1.6, and 4.2 percent increase in large-scale incident durations, respectively. In addition, longer response times for towing and highway patrol are also significantly associated with longer incident durations. Given large-scale incidents, incidents involving vehicle fire or unscheduled roadwork are likely to last longer on average. Finally, large-scale incidents on weekdays, afternoon peaks, and on higher AADT roads last shorter, however the magnitude (and in some cases direction) of associations is heterogeneous, i.e., the direction can be positive in some cases and negative in other cases. The practical implications of results for large-scale incident management are discussed.
Driver Expectations and Understanding of Wrong-Way Driver Warning Messages Displayed on Dynamic Message Signs
Melisa Finley, Texas A&M Transportation InstituteShow Abstract
Nada Trout, Texas A&M Transportation Institute
While extensive human factors and traffic operations research on dynamic message sign (DMS) message design has been previously conducted, these efforts have not looked at the design of wrong-way driver (WWD) warning messages. In 2013, researchers conducted focus groups to obtain motorists’ input regarding the design of WWD warning messages. Researchers also reviewed previous literature and DMS message design manuals to gain insight into the design of WWD warning messages. Based on the findings from these efforts, researchers developed a single-phase WWD warning message that can be displayed on DMSs when a WWD is reported. Researchers also developed an alternative single-phase WWD warning message for DMSs with only 15 characters per line.
In 2015, researchers conducted motorist surveys to assess motorist understanding of these WWD warning messages. These survey findings supported the use of the two WWD warning messages. However, there was some evidence that the abbreviation for vehicle (VEH) when used with WRONG WAY may be initially misunderstood. Therefore, the alternative message should be used sparingly.
These WWD warning messages should be posted on DMSs whenever a WWD is reported, even if not confirmed, in order to alert motorists to the possibility of a WWD as soon as possible. Anytime a WWD warning message is displayed on a DMS, the beacons located on the DMS should be activated. If the DMS does not have beacons, the entire message may be flashed. One line of these messages should never be flashed.
Twitter or Chatter? Involving Social Media Data Analysis in Traffic Incident Management
Ying Chen, Northwestern UniversityShow Abstract
Archak Mittal, Ford Motor Company
Hani Mahmassani, Northwestern University
The contents posted by users on social media sites generate large volumes of data. The feasibility of using social media data, specifically from Twitter, for detecting traffic incidents is evaluated. For the purpose of incident management, a text-mining process is presented to extract and search real-time traffic-related Twitter data by two methods, keywords search and specific users search. The presented framework consists of three main components, Twitter data mining, location extraction and traffic management. The approach is implemented using data from the Chicago area; using on-line simulation-based traffic estimation and prediction tools, traffic management strategies developed to reduce the impact of incidents are evaluated. The results confirm the potential of Twitter posts to complement and improve the effectiveness of dynamic traffic management approaches for incident conditions.
Heterogeneity in Incident Durations: Comparison of Random Parameter and Quantile Regressions
Behram Wali, Urban Design 4 Health, Inc.Show Abstract
Asad Khattak, University of Tennessee
Jun Liu, University of Alabama
Accurate prediction of incident duration and response strategies are two imperative aspects of traffic incident management. Past research has applied various types of regression models for predicting incident durations and quantification of associated factors. However, an important methodological aspect is unobserved heterogeneity which may be present due to several unobserved/omitted factors. Incorporation of heterogeneity has significant potential to enhance predictive capabilities as well as obtain more robust insights for designing practical incident management strategies. This study uses two statistical techniques, random parameter and quantile regression models, to address unobserved heterogeneity and to compare both methodologies with respect to important aspects of incident management. By using 2015 Virginia incident data related to more than 45,000 incidents, the associations of incident durations with several factors including detection source, incident type, roadway type, temporal factors, and incident characteristics are explored. Specifically, as the name implies, quantile regression models associations between different quantiles of incident duration and explanatory factors. This facilitates designing appropriate strategies for small, medium and large-scale incidents. Compared to quantile regression and fixed parameter models, random parameter models can potentially give more accurate predictions of incident durations. However, they do not (typically) capture different quantiles of incident durations. The practical implications of results are discussed from the perspectives of travelers and incident managers.
Shockwave Analysis of Freeway Congestion
Yibing Wang, Zhejiang UniversityShow Abstract
Yuheng Kan, Zhejiang University
Ludovic Leclercq, Université Gustave Eiffel
Shockwave analysis is crucial for a variety of tasks in traffic engineering. There are some deficiencies with available shockwave analysis results. Firstly, simple cases were mostly considered with qualitative shockwave analyses. Secondly, field data is generally scarce for evaluating shockwave dynamics. Also, the ad hoc availability of field data limits the possibility of systematic and thorough examination of shockwave formation and propagation in genuine situations. Thirdly, although ramp metering is an important means applied to prevent congestions on freeways, available studies on ramp metering was rarely accompanied with shockwave analysis. This paper conducts qualitative shockwave analyses and quantitative evaluation with respect to a typical traffic scenario on freeways. The evaluation results obtained via macroscopic traffic flow simulations and field data testing were found to match the qualitative results of shockwave analyses quite well.
Characterizing and Ranking Recurring Freeway Bottlenecks
Ishtiak Ahmed, North Carolina State UniversityShow Abstract
Nagui Rouphail, North Carolina State University
Shams Tanvir, University of California, Riverside
Lu Pan, University of Southern California
This study proposes a method that can identify recurring freeway bottlenecks and estimate their impacts. The bottleneck identification method is based on a modification of previous research. Probe vehicle speed data are used to signal bottleneck activation on a road segment at a particular clock time and day. The impact of daily congestion originating from a bottleneck is calculated to characterize its variability and weekday patterns. The temporal probability of bottleneck activation is estimated in order to filter recurring bottlenecks. Finally, the probability of bottleneck activation, upstream queue length, duration, and severity of the resulting congestion were integrated into two bottleneck impact performance measures.
A case study conducted on I-40 in North Carolina using October 2015 data was carried out to demonstrate the proposed method. The analysis led to the identification of ten recurring bottleneck regions in a 422 centerline miles of the facility. The impacts of those bottlenecks using the proposed approach were compared to two previously established bottleneck identification and ranking by the VPP project, and the State of Florida. Variations between methods were detected, and the contributing factors to those differences and their limitations are also discussed.
Keywords : Freeway, Congestion, Recurring bottleneck, Queue length, Ranking
Impact of Precipitation on Freeway Free-Flow Conditions: Exploratory Analysis of Time Sensitivity
George Gillette, Texas A&M Transportation InstituteShow Abstract
Kay Fitzpatrick, Texas A&M Transportation Institute
Raul Avelar, Texas A&M Transportation Institute
Understanding the impact of inclement weather on freeway conditions has been an ongoing effort for many years. The impact of precipitation upon traffic characteristics varies largely between studies; estimates for the reduction in free-flow speed by light rain range from 2 to 13 percent. Many of these differences can be easily attributable to the unique site characteristics of each study, and it is thereby critical to generate analyses at a wide variety of locations.
This study evaluates the impact of a rainfall event on the free-flow condition of a freeway, and aims to provide an additional dimension to the analysis by understanding how the length of the rainfall event impacts the driver speed behavior. Hourly intervals of speed and volume data were collected from archived reports from 20 unique segments in California, and correlated to three years of precipitation data from four weather stations. Over the course of a rainfall event, it was found that both the volume and speed on the freeway is reduced. From a random-effects model constructed with the subset of daytime data, the number of hours of the rainfall event was found to reduce free-flow speed by 0.42 mph per hour, controlling for the other predictors. The model estimates that the impacts of light and moderate rain conditions are approximately 3 and 10 percent reduction to free-flow speed. These results are slightly above the reported HCM 2010 values of 2.01 and 7.24 percent respectively, and emphasize the importance of location-specific analyses.
Long-Range Dependence of Traffic Flow and Speed of a Motorway: Dynamics and Correlation with Historical Incidents
Sai Chand, University of New South WalesShow Abstract
Gregory Aouad, WSP
Vinayak Dixit, University of New South Wales
Speed and flow of vehicles tend to have several effects on the dynamics of a transport system. Fluctuations of these variables can implicate congestion, lower predictability and may even catalyze crashes. This paper utilizes a concept of fractal theory called the Hurst exponent, a measure of the long-range dependence (LRD) of a time-series to understand the fluctuations in flow and speed of a motorway in Sydney, Australia. The spatial and temporal variation of LRD for flow (Hflow) and speed (Hspeed) at several monitor sites were discussed. Furthermore, the effects of number of lanes on flow and speed predictability were also explored. It was observed that the flow predictability of 2-lane sections is significantly lower when compared with 3-lane and 4-lane sections. Conversely, the speed predictability of 4-lane sections is significantly higher than 2-lane and 3-lane sections. Finally, traffic congestion was defined in terms of LRD of speed and its correlation to historical incident rates was measured. It was ascertained that monitor sites with a historically high proportion of large Hspeed, were correlated with unsafe locations. It is anticipated this study would lead to many applications of fractal analysis on highways and urban traffic.
Improving Traffic Incident Management in Virginia: Research Motivated by Public- and Private-Sector Perspectives
Lance Dougald, Virginia Department of TransportationShow Abstract
Noah Goodall, Virginia Department of Transportation
Ramkumar Venkatanarayana, Virginia Transportation Research Council
Traffic congestion and safety on U.S. roadways are issues that receive significant and continual attention from transportation professionals. Improving traffic incident management (TIM) is one means to help reduce congestion, as traffic incidents account for approximately 25 percent of total congestion on U.S. highways. The purpose of this research was to investigate TIM initiatives including quick clearance practices and policies used by other transportation agencies, to identify quick clearance barriers in Virginia, and to assess the feasibility of adopting strategies that are not currently implemented in the Commonwealth. The research tasks included a review of best practices, an analysis of the Code of Virginia, and 36 interviews with TIM stakeholders from Virginia including the department of transportation, fire, police, emergency dispatch, and private towing companies. Interviewees reported numerous factors contributing to slow incident response and clearance. The paper describes initiatives adapted in other states or suggested by interviewees such as stronger authority removal and abandoned vehicle legislation, innovative towing incentive and compensation programs, and technologies and policies to accelerate crash investigations. The results of the research led to the implementation of pilot projects for evaluation. The methodology and findings should be relevant to other transportation and emergency management agencies with TIM responsibilities.
Combined Variable Speed Limits and Ramp-Metering System for Capacity Drop Control at Merge Bottleneck
Hyun Woong Cho, Georgia Institute of Technology (Georgia Tech)Show Abstract
Jorge Laval, Georgia Institute of Technology (Georgia Tech)
This paper proposes a variable speed limit and ramp metering (VSL-RM) control strategy to prevent and recover from capacity drops at freeway merge bottlenecks. Using kinematic wave theory, we derive analytical models and time-space diagrams and present micro-simulation models. We find that the combed VSL-RM system outperforms either component in isolation for preventing or curtailing traffic breakdown; if only one component has to be used, its choice depends on the distribution of traffic demand. The proposed strategy can be implemented with current technology.
Effects of Rain on Freeway Traffic in Southern California
Sawanpreet Singh Dhaliwal, California State Polytechnic University, PomonaShow Abstract
Xinkai Wu, California State Polytechnic University, Pomona
John Thai, City of Anaheim
Xudong Jia, California State Polytechnic University, Pomona
A number of studies in the past have quantified the impact of rain on traffic parameters but all of them were limited to wet areas. The research reported in this paper have expanded the literature by studying the impact of rain in a dry area such as Southern California, and investigating for regional differences in the impact. Traffic data (loop detectors) and precipitation data (rain gauges) from Los Angeles Metropolitan Area has been analyzed to access the impact of rain on traffic stream parameters such as free-flow speed, the speed at capacity, and capacity. Rainfall events have been categorized as light, medium, and heavy as discussed in Highway Capacity Manual (HCM) 2010. Density plots and fundamental diagrams for different rain types proved that free-flow speed, speed at capacity and capacity are reduced by 5.7%, 6.91%, and 8.65% respectively for light rain, 11.71%, 12.34%, and 17.4% respectively for medium rain, and 10.22%, 11.85% , and 15.34% respectively for heavy rain. The reductions for free-flow speed are lower, whereas, for speed at capacity and capacity are higher than those reported in HCM 2010. Moreover, the headway is increased during rain that exhibits cautious driving behavior. Multiplicative weather adjustment factors have been computed to compensate the loss of speed and capacity. This paper also demonstrates the spatial and temporal impact of rain on traffic. Downstream traffic is not much affected by a rainfall event while the upstream traffic is negatively impacted. This paper is expected to support weather-responsive traffic management strategies for dry areas.
Using Temporal Detrending to Observe the Spatial Correlation of Traffic
Alireza Ermagun, Northwestern UniversityShow Abstract
Snigdhansu Chatterjee, University of Minnesota, Twin Cities
David Levinson, University of Sydney
This empirical study sheds light on the correlation of traffic links under different traffic regimes. We mimic the behavior of real traffic by pinpointing the correlation between 140 freeway traffic links in a sub-network of the Minneapolis - St. Paul freeway system with a grid-like network topology. This topology enables us to juxtapose positive correlation with negative correlation, which has been overlooked in short-term traffic forecasting models. To accurately and reliably measure the correlation between traffic links, we develop an algorithm that eliminates temporal trends in three dimensions: (1) hourly dimension, (2) weekly dimension, and (3) system dimension for each link. The correlation of traffic links exhibits a stronger negative correlation in rush hours, when congestion affects route choice. Although this correlation occurs mostly in parallel links, it is also observed upstream, where travelers receive information and are able to switch to substitute paths. Irrespective to the time-of-day and day-of-week, a strong positive correlation is witnessed between upstream and downstream links. This correlation is stronger in uncongested regimes, as traffic flow passes through consecutive links more quickly and there is no congestion effect to shift or stall traffic. The extracted correlation structure can augment the accuracy of short-term traffic forecasting models.
Freeway Traffic Queue-Warning System: Deployment and First Impressions
Peter Dirks, University of MinnesotaShow Abstract
Zhejun Liu, University of Minnesota, Twin Cities
John Hourdos, University of Minnesota, Twin Cities
The formation and propagation of queues and traffic shockwaves on urban freeways is an unavoidable result of the ever increasing demand for transportation. Beyond the spread of congestion and reduction of system productivity, some of these traffic patterns can cause serious rear-end and other types of collisions. This paper discusses the deployment of a prototype, infrastructure based, Queue Warning system that detects crash-prone conditions in freeway traffic and provides warning messages to drivers. The system, by raising the alert level of the drivers aims in reducing the frequency of crashes. A review of existing queue-warning systems is included as well as a description of the basic inner workings of the system. A description of the deployment site is followed by the results of an initial evaluation of the currently operating Queue Warning system.
Queue Length Estimation for Freeway Facilities Based on the Combination of Point Traffic Detector Data and Automatic Vehicle Identification Data
Somaye Fakharian Qom, California Department of Transportation (CALTRANS)Show Abstract
Mohammed Hadi, Florida International University
Yan Xiao, Florida International University
Haitham Al-Deek, University of Central Florida
Queue length is a critical performance measure in the assessing and managing of transportation network performance. This study explores the methods to estimate queue length based on point traffic detector and Automatic Vehicle Identification (AVI) data. Two new methods that integrate data from point traffic detectors and automatic vehicle identification (AVI) readers to estimate the queue length of freeway segments for both off-line and real-time applications are developed in this study. In the first method, the queue length between two detectors is estimated using linear interpolation between the travel time measurement based on AVI data, when the link is fully queued and when no queue is present. In the second method, a segment with a partial queue is divided into two subbasements; the first subsegment is assumed similar to the upstream traffic conditions, and the second one to the downstream traffic conditions. Then, the length of each part is calculated based on AVI speed data.
The performances of these methods are assessed and compared in two case studies based on simulation data and real-world data. The results show that utilizing a combination of point detector data and AVI data produces accurate estimates of queue length. The queue estimation method based on cumulative volumes collected using point detectors alone can also produce reasonably good estimates but requires additional ramp detection and assumptions regarding moving queue density. The two combination methods produce results that are close to each other based on the simulation data and real-world data. The segmentation method produces better results based on the real-world data.
Keywords: Queue Length, Point Traffic Detector, Automatic Vehicle Identification, Simulation, Real–World ITS Data.
Identifying Wrong-Way Driving Hotspots by Modeling Crash Risk and Analyzing Traffic Management Center Response Times
Adrian Sandt, University of Central FloridaShow Abstract
Haitham Al-Deek, University of Central Florida
John H. Rogers, Jr., University of Central Florida
Since wrong-way driving (WWD) crashes are often severe, it is important for transportation agencies to identify WWD hotspot segments that deserve priority WWD countermeasure implementation. Two different approaches to identify these hotspot segments were developed and applied to a case study, Central Florida limited access highways. The first approach uses a Poisson regression model that predicts the number of WWD crashes in a roadway segment based on WWD citations, 911 calls, traffic volumes, and interchange designs. Combining this predicted crash value with the actual number of WWD crashes in the segment gives the WWD crash risk of the segment. Ranking roadway segments by WWD crash risk provides agencies with an understanding of which segments have high WWD crash frequencies and high potential for future WWD crashes. This approach was previously applied to South Florida, while the present research paper extends this approach to the Central Florida area. The second approach is based on operational data collected in traffic management centers and can be used if accurate WWD 911 and citation data with GPS location are not available or as a supplement to the first approach. It identifies and ranks WWD hotspots based on the time spent responding to WWD events. In Central Florida, the results of the above two approaches agreed with each other, showing that these approaches can be used independently or jointly. Together, these approaches can help transportation agencies determine the regional WWD hotspots so they can cooperate and implement WWD countermeasures at these locations.