2020

Chang-Jen Lan, Town of Jupiter

The queue percentile formulation adopted in the 2010 Highway Capacity Manual (HCM) is empirical based.  It contains an overly simplified functional form and parameters such that the accuracy of the back-of-queue percentile estimates is compromised.  In view of this, the author applies the Delta method to derive the uncertainty formulation of back-of-queue through the first degree of Taylor expansion approximation, assuming that the source of uncertainty comes from arrival flow, saturation flow and, for actuated controller, green time variability from vehicle actuation.  Due to the fact that multiplication of the independent, positive random variables results in a lognormal distributed outcome, the back-of-queue percentile is quantified as a lognormal predictive interval function.  The simulation study shows that the proposed back-of-queue predictive interval is much more accurate than the HCM counterpart in almost all the cases evaluated based on the Kolmogorov–Smirnov goodness of fit test.  The proposed formulations can be used to quantify back-of-queue predictive interval for both pretimed and actuated control, while HCM is indifferent to the type of control.  As expected, the variability of back-of-queue from the actuated control is less than the pretimed control in unsaturated condition.  The predictive interval profiles become identical when the degree of saturation approaches to unity and maxes out the green time.

Weike Lu, Soochow University
Jun Liu, University of Alabama
Jiannan Mao, Southwest Jiaotong University
Guojing Hu, Jackson State University
Chuqiao Gao, Southwest Jiaotong University
Lan Liu, Southwest Jiaotong University

Evaluating a regional traffic control system often requires understanding of both advantage and disadvantage of control schemes as well as the interrelated characteristics of the system. Based on the Macroscopic Fundamental Diagram (MFD) concept, this study proposes four indicators, including maximum throughput, critical accumulation, gridlock accumulation and homogeneousness degree, to assess the efficiency of regional signal control schemes in a road network. The maximum throughput and gridlock accumulation can be used to reflect road network capacity and load capacity, respectively. The homogeneousness degree quantifies the spatial variations of traffic flows in the network. The critical accumulation combined with the gridlock accumulation values the durability of a regional control system in managing congestions in the network. This study used the regional road network in Qingyang District of Chengdu, China as a real-world example to demonstrate the proposed MFD-based approach. In the demonstration, the MFD-based evaluation method was compared to traditional travel time-based method. The demonstration evaluated the control effect and characteristic values of the network under four control modes including fixed-time, actuated, adaptive and adaptive coordinated control.

Rui Yue, Beijing Jiaotong University
Guangchuan Yang, Institute for Transportation Research and Education
Dongmei Lin, DKS Association
Zong Tian, University of Nevada, Reno

Traffic signal retiming is a cost-effective signal timing approach that periodically upgrades exiting signal coordination plans of signalized corridors. Nevertheless, there are still several limitations with existing signal retiming methodologies, such as managing signal data, optimization of coordination parameters, diagnosis of timing errors, visualization of signal performance. To fill this gap, this research introduces a novel signal timing approach, which offers local agencies an alternative way to assess and improve the operation of existing signalized corridors. It allows for displaying the vehicle trajectories in the time-space diagram (TSD), based on which, transportation engineers could diagnose the potential issues with a signal coordination plan and develop an optimized signal coordination plan for the corridor. In addition, a unique signal performance index (SPI), which took into account average travel speed, number of stops, cycle length and segment length adjustments, were developed to visualize the performance of a signal coordination plan. A total of 36 signal coordination plans were applied on seven signalized corridors in the City of Reno, Nevada based on the proposed signal retiming approach. Through a before-after study, this research concluded that traffic signal retiming considerably improved the travel speed of the corridors and reduced delay at signals. SPI level of 75% of the retimed signal coordination plans was increased, and 17% of the retimed plans witnessed increases in either speed score or stop score. These case studies proved that the proposed signal retiming approach is applicable for field practices.

Huihui Lan, Beijing Jiaotong University
Xianyu Wu, Beijing Jiaotong University

In order to accurately divide the traffic control sub-area and improve the regional coordination control effect, this paper studies the key technologies of traffic control sub-area division from the influence of the discrete characteristics of the fleet on the coordinated control. Correlation analysis method and regression analysis method are used to construct the correlation model of adjacent intersections. The principle and process of transportation sub-area division are designed by using analytic hierarchy process, and a new traffic sub-area division model is proposed. Taking Gulou Street in Miyun District of Beijing as an example, combined with the green wave coordination optimization method, the new model was applied and simulated in SimTraffic. The simulation results show that the average delay of the intersection is reduced by 21.4%, the total number of stops is reduced by 9.62%, and the travel time is shortened by 8.75%. The new intersection degree model and the coordinated control scheme obtained by the traffic sub-area division model make the highest traffic efficiency. That is to verify the scientific research method of this paper, the research results are reliable, and the practical application value is high.

Hatem Abou-Senna, University of Central Florida
Essam Radwan, University of Central Florida
Hesham Eldeeb, University of Central Florida

The four-section head Flashing Yellow Arrow (FYA) provided an opportunity to advance the operation of the left turn mode at intersections. The UCF team developed an exclusive hardware platform. The hardware platform was developed for two main objectives: first, to provide a generic device compatible with the different controller types used by different jurisdictions within the FDOT Districts and, second, to automate selection of the FYA left-turn modes based on available gaps in the opposing traffic at intersections acquired in real time from existing sensors in the field. Phase III provided conclusive offline testing using online and field testing of the DSS. The testing verified that the system was able to accurately acquire real-time traffic conditions and react by switching between red and FYA modes in a rational manner consistent with driver expectations and left-turning gap acceptance thresholds. It was also concluded that coordinated signals with very long cycle lengths, 3 minutes and longer, help in providing adequate gaps even in heavy traffic patterns since most of the vehicle arrivals are in platoons and at the beginning of the cycle. However, short cycle lengths eliminate sufficient gap times even with coordination. Two methods were used to calculate the minimum gap: discrete and average approach. Although the average method provides a more conservative approach, the discrete approach was more accurate. The developed platform is applicable at any four-section head configuration to alter the left turn restriction throughout the day to maximize safety and efficiency of the intersections.

Min-Wook Kang, University of South Alabama
Moynur Rahman, City of Mobile
Joyoung Lee, New Jersey Institute of Technology

The objective of the present study is to see how the dilemma zone length and location are related to intersection safety at rural high-speed signalized intersections. For that, dilemma zones at thirty rural signalized intersection approaches where similar traffic operations and land-use conditions are present were identified, and traffic conflicts associated with the dilemma zones were investigated. Divers’ stop or go behaviors as well as vehicle dynamics (e.g., vehicle speeds and locations) during the yellow and red clearance intervals were collected and analyzed to determine the dilemma zone length and location of each site. Red-light running (RLR) violations and vehicle’s abrupt stops were also collected to investigate the relationship between such conflicts and dilemma zone length and location. As a result, two dilemma zone conflict models were developed. The analysis results show that the conflict models are accurate enough to predict the safety level of high-speed signalized intersections using the two dilemma zone variables. Results show that the chance of intersection angle conflicts increases if the dilemma zone is located farther from the intersection stop bar. Results also show that there would be a high chance of rear-end conflicts if the dilemma zone length is longer. The models were validated with additional datasets, and acceptable RMSE and MAPE values were obtained as a result.

Montasir Abbas, Virginia Polytechnic Institute
Charles Nichols, Virginia Polytechnic Institute
Anne Elise Thomas, Virginia Polytechnic Institute
Qichao Wang, National Renewable Energy Laboratory (NREL)

The goal of this research is to improve signal operation and reduce delay through an innovative transdisciplinary integration of traffic signal control and music theories. The main hypothesis in this research is that “the underlying sense that musicians exhibit during improvisation can be utilized to better synchronize signal control with internal traffic rhythm.” In this work, we developed and used a novel Signal Operation with Neuro-fuzzy Acoustic Tuning Application (SONATA) interface to control a simulated traffic signal. Three musicians utilized the SONATA framework to control their respective traffic signal using visual/auditory information and auditory only information to minimize queues. It was found that two of the three musicians in this study have outperformed the Webster’s optimal control (by 22.3% and 17.6%). It was also found that all three musicians controlled the traffic signal better when they were not looking at the simulation (i.e., using only auditory sensing). This paper lays out the foundation for a new area of research and proves that harmonization of control and traffic with music-based control has merits.

Hao Liu, Pennsylvania State University
Randy Machemehl, University of Texas, Austin

This paper models a signalized intersection as a queueing system with one server and multiple channels and proposes a stochastic traffic delay model that can be used for an adaptive signal control method. We assume the arrivals of each approach can be described as Poisson processes and the departure rates are constant during the green time. Based on queueing theory, we develop the probability distribution function (pdf) of the busy period of a M/D/1 queue given the initial queue length. Then, the expectation of traffic delay incurred in this busy period is derived. Founded on these results, a cycle-based stochastic traffic delay model is proposed. Further, by utilizing the rolling horizon scheme, we propose an adaptive signal control model minimizing the cycle-based traffic delay per vehicle and demonstrate its efficiency by comparing to Webster’s model through simulation. The proposed model is beneficial to adaptive traffic signal control methods and policy design of general polling systems.

Hao Liu, Pennsylvania State University
Amber Chen, University of Texas, Austin
Randy Machemehl, University of Texas, Austin

An adaptive signal control framework for a single intersection is developed in this paper. Traffic volume prediction and signal optimization models are two crucial components of adaptive control methods. This paper uses the Barron-Jensen/Frankowska (B-J/F) solution to the Lighthill-Whitham-Richards (LWR) model to solve the traffic volume prediction problem as Linear Program (LP). Predicted volumes are used as inputs in a mixed integer linear program (MILP) that finds optimal phase sequence and splits within a pre-defined project horizon. The B-J/F solution is an analytical and grid-free solution of the LWR model, and it enables us to derive the constraints without the restriction on the type of the boundary condition, such as uniform inflows or Poisson arrivals. This model is applicable rapidly changing traffic conditions, since it depends on the current traffic state of the network instead of assumptions about the arrival process and the correlation of flows in the time dimension. This paper also presents an approximate method to reduce the computational time of the proposed MILP, which allows us to find optimal solutions within 0.05 seconds. In a practical case study, the solutions obtained using the proposed model improve upon those generated by the critical lane flow ratio method. Delay reductions between 13.9-27.2% are observed for a range of traffic congestion states.

Haoxuan Fan, Southeast University
Gang Ren, Southeast University

Considering the coordinated optimization of the signal timing of intersections on trunk road, a multi-objective function is formulated by choosing vehicle delay, vehicle stop and exhaust emission as optimization goals. Then a fuzzy compromise programming approach is employed in this paper and exponential membership function is used in this approach. Two typical adjacent intersections of a trunk road in Hefei city are selected as study case. By using particle swarm optimization algorithm, the optimized signal timing scheme can be obtained. The results show that coordinated optimization of the signal timing of intersections on trunk road has an optimal effect on reducing vehicle delay, vehicle stop and exhaust emission.

Yujing Zheng, Tongji University
Yu Shen, Tongji University
Yuxiong Ji, Tongji University
Yuchuan Du, Tongji University
Xiaoyu Cai, Chongqing Jiaotong University

Conflicts between crossing pedestrians and right-turn vehicles not only bring a safety concern, but also reduce the right-turn capacity, leading to queue spillback when the right-turn vehicle volume is large. We develop a signalization approach to improve the right-turn capacity. The developed approach terminates the pedestrian green signal early to provide a conflict-free interval to facilitate the movements of the right-turn vehicles. The negative impacts of the early termination on pedestrians are expected to be small since the volume of crossing pedestrians is usually low at the late stage of the crossing phase. The developed approach balances the priority needs between improving the right-turn capacity and reducing the negative impacts on pedestrians. It determines the termination time in real time based on vehicle queue length and pedestrian volume on the crosswalk. We validate the developed approach numerically based on actual data at an intersection in Shanghai, China. The results demonstrate that the developed approach efficiently reduces vehicle queue lengths with limited impacts on pedestrians.

Yongtao Zheng, Southeast University
Xuedong Hua, Southeast University
Wei Wang, Southeast University
Jialiang Xiao, Southeast University
Dongya Li, Southeast University

Facing the signalized intersection with quantitative unbalance between approach and exit lanes, this study proposes a method of dynamic use of left-turn lane for opposite through, to improve the efficiency of signalized intersection. The complete scheme of this method mainly consists of three parts: the location routes of dynamic lane, the temporal scheme and the spatial scheme. To demonstrate the effectiveness of this method, a simulation scheme comparing the intersection with and without the dynamic lane is designed, using VISSIM. The parameters of simulation model, including Traffic simulation environment parameters and VISSIM build-in models, are calibrated by the field data collected from an intersection in Nanjing, P. R. China. The results of simulation indicates that this method decreases average delay and average queue length of intersection effectively for the intersection with quantitative unbalance between approach and exit lanes, under the conditions of different through volumes or left-turn ratios. The maximum can be 46.3% (average delay) and 69.2% (average queue length) for all situations. At last, three suggestions are presented for the application of this method.

Hao Yu, Southeast University
Jie Bao, Nanjing University
Lu Bai, Southeast University
Pan Liu, Southeast University

This study evaluates the performance of three different traffic signal control strategies in normal and incident conditions. The three signal control strategies are the fixed signal control, adaptive signal control minimizing vehicle delay, and adaptive signal control maximizing intersection capacity. The performance of the signal control strategies is evaluated in a simulation environment. The double queue model is applied in the evaluation framework to capture the traffic dynamics and possible queue spillbacks. Three levels of traffic demand are considered, including the low, medium, and high traffic demand scenarios. The simulated network performance is compared between normal and incident conditions considering three traffic signal control strategies. Three performance measures are considered in the comparative analysis, including the system total travel time, the difference in system travel time between incident and normal conditions, and diagram of network throughput-vehicle number. Based on the results of comparative analyses, it is concluded that the ASC-IC, which focuses on maximizing intersection capacity, outperforms other signal control strategies in heavy traffic conditions. When traffic demand is low or median, priority should be given to the ASC-VD, which focuses on minimizing vehicle delay. Even though the FSC provides the best stability in incident conditions, the strategy is not recommended considering the lowest system efficiency. Future research in more realistic conditions is warranted.

Muriel Verkaik-Poelman, Delft University of Technology
Andreas Hegyi, Delft University of Technology
Alexander Verbraeck, Delft University of Technology
Hans van Lint, Delft University of Technology

Signalized traffic control is important in traffic management to reduce congestion problems in urban regions. With recent technological developments, more data becomes available to feed the controllers and advanced state estimation and prediction methods are developed based on these data. To fully benefit from these techniques in the design of signalized traffic controllers, it is important to look at the quality of the estimated and predicted input quantities in relation to the performance of the controllers. Therefore, in this paper, a general framework for sensitivity analyses is proposed, to analyze the effect of erroneous input quantities on the performance of different types of signalized traffic control. The framework is illustrated for phase-based predictive control. Experimental relations between the performance of the control system and the prediction horizon are obtained. Showing an increase in performance for increasing prediction horizons, even in most of the cases with erroneous input data. These experimental relations are studied for different type of controllers with increasing degrees of freedom, finding a trade-off between sensitivity and adaptivity of the controller. The outcome of the sensitivity analysis contributes in understanding the relations between information quality and performance of signalized traffic control. Moreover, in the design phase of controllers, sensitivity analysis results can be used to make choices on the length of the prediction horizon, the level of adaptivity of the controller, the objective of the control system, and the input quantities that need to be estimated and predicted the most accurately.

Michael Adamson, Fehr & Peers
Grant Schultz, Brigham Young University
Mitsuru Saito, Brigham Young University
Michael Stevens, Brigham Young University

The purpose of this research was to evaluate the safety and operational differences among permitted, protected, and protected-permitted left-turn phasing. Permitted phasing allows vehicles to turn left after yielding to opposing vehicles; protected phasing provides an exclusive phase for vehicles to turn left that does not allow opposing vehicles; and protected-permitted phasing combines the previous phasing alternatives, allowing both permitted and protected turning movements. Intersections with 1-, 2-, and 3-opposing-lane configurations with permitted and protected-permitted models (split into green times of 10-, 15-, and 20-seconds) were modeled for a total of 12 simulation models. Each model was divided into 100-225 different volume scenarios, with incremental increases in left-turn and opposing volumes. By exporting trajectory files from VISSIM and importing these into the Surrogate Safety Assessment Model (SSAM), crossing conflicts for each volume combination in each model were identified and extracted. MATLAB was then used to create contour maps; the contours represent the number of crossing conflicts per hour associated with different combinations of left-turn and opposing volume. Basic decision boundaries were examined in each contour map. JMP Pro statistical analysis software was used to perform a linear regression analysis on transformed data and to develop natural log-based equations that would form the decision boundaries for each configuration and phase alternative. These equations were then graphed and final decision boundaries were developed for the 1-, 2-, and 3-lane configurations between permitted and protected-permitted phasing as well as between protected-permitted and protected phasing.

Junwoo Song, Imperial College London
Ke Han, Southwest Jiaotong University
Simon Hu, Zhejiang University

Real-time traffic management with environmental objectives has been a difficult challenge due to (1) the highly dynamic and uncertain nature of road traffic and their emission profile; (2) the need for generating timely and robust decisions for large-scale networks; (3) the incorporation of potentially massive and heterogeneous on-line data; and (4) the balance between traffic and environmental objectives. To address these challenges, this paper proposes a real-time traffic signal control framework based on a nonlinear decision rule (NDR), encapsulating an two types of neural networks (such as FFNN and RNN), which is integrated with realistic traffic and emission modelling via agent-based micro-simulation. The implementation of the NDR method is divided into an off-line module and an on-line module. The off-line module trains the neural network (FFNN or RNN) and the NDR using historical traffic data, which amounts to a stochastic optimization with traffic and environmental objectives. The on-line module efficiently implements the trained NDR, whose performance is guaranteed by the off-line training. Simulation study of the proposed signal control framework in a real network in West Glasgow demonstrates the advantage of the proposed method over the current signal control parameters in terms of alleviating congestion and reducing traffic emissions.

Nemanja Dobrota, Kittelson & Associates, Inc.
Rodolfo Alfaro-Carcoba, University of Nevada, Reno
Nikola Mitrovic, CHA Consulting, Inc.
Aleksandar Stevanovic, University of Pittsburgh

Traffic conditions on urban arterials vary significantly over time, exhibiting variations in traffic volumes. Appropriate scheduling of multiple (time-of-day) signal timing plans is often seen as an efficient and cost-effective solution, in coping with inevitable fluctuations. Although the abundant volume of data has become recently available, most of the agencies still continue to use only small datasets (containing up to several weeks of data) in optimizing and scheduling signal plans. In this paper we propose an easy-to-apply data mining approach that can help agencies evaluate and (potentially) improve the existing scheduling. The proposed method uses K-means clustering to identify the dominant traffic profiles throughout a year along a $3.6$-mile section of coordinated arterial in Fort Lauderale, FL. To select appropriate number of clusters the elbow curve, silhouette method and design traffic (or centroid vector) are considered. Then, for each $30$-min interval within a week the appearance / frequency of all identified clusters is computed and plotted. Following the Visual Analytics [VA] principles, that integrates domain and theoretical knowledge, we aggregate certain clusters until the distinctive time-of-day patterns are revealed. The proposed method reveal certain dissimilarities with the deployed scheduling. In particular the Friday traffic is quite distinctive and quite different from other weekday traffic while the Saturday’s midday traffic is much higher than traffic on Sunday although the single signal timing plan is deployed over the weekend. As proposed method shown a great potential for identification TOD breakpoints, the next step should be validation in microsimulation for longer time periods.

Yen-Hsiang Chen, National Taiwan University
Yao Cheng, University of Maryland, College Park
Gang-Len Chang, University of Maryland, College Park

Most congested urban arterials, connecting multiple major intersections with heavy turning flows, are likely to comprise multiple dominant path flows in addition to its one- or two-way through traffic. Hence, those arterials, even with a signal plan designed to maximize progression, often suffer from mutual blockage between the through vehicles and turning flows. This study proposes a model for design of a progression band for each candidate path along the target arterial with only each intersection's volume counts. The proposed model will construct local progression bands for vehicles on all local paths, constituted by a pair of each link’s upstream flow-in and downstream flow-out movements. Such identified local progression bands are then designed to optimally connect two neighboring links under the objective of maximizing the total weighted progression bands.   To ensure the effectiveness of the produced signal plan, the proposed model can exercise its embedded functions to estimate some key information associated with design of offsets and phase sequences, including the impeded duration within a local path's progression band due typically to mutual queue blockage, and estimated vehicle volume within each designated progression band. With extensive simulation experiments based on an actual arterial and field data, the evaluation results confirm that the proposed model can indeed produce progression bands for all local and multi-link path flows, and yield the overall optimized state for the arterial with the MOEs of total weighted progression bands, number of vehicle stops, and the total traffic delay.