2019

Yao Cheng, University of Maryland, College Park
Hyeonmi Kim, University of Maryland, College Park
Gang-Len Chang, University of Maryland, College Park

Despite the extensive studies aiming at contending with congestion on urban arterials, an effective model to produce optimal signal progression for an arterial experiencing both heavy bus and passenger car flows remains unavailable. In response to such needs, this study has presented a signal optimization model that can offer concurrent progression to both modes or to a selected mode(s) in a selected direction(s), based on traffic volume, bus ratio, and geometric conditions. To capture the operational features of both modes, the proposed model has effectively taken into account all critical issues that may result in mutual impedance between these two modes, which include the potential blockage of passenger car queues to the roadside bus stops, the excessive start-up delays caused by transit vehicles queuing at the intersection stop line, and the reduced travel lanes for progressing flows due to buses dwelling at roadside stations with limited storage capacity. In addition, by weighting the bandwidths with the passenger volumes by mode and by direction, the proposed model is capable of offering the progression only to the mode(s) and the direction(s) that are justified to do so from the perspective of maximizing the benefits for all the arterial users. Our numerical analysis results have confirmed the effectiveness of the proposed model in producing concurrent progression bands for both modes under various realistic constraints and volume levels. Further evaluation with extensive simulation experiments has also demonstrated that the benefits offered by the proposed model will not be at the cost of other MOEs.

Marija Ostojic, Northwestern University
Archak Mittal, Ford Motor Company
Hani Mahmassani, Northwestern University

Vehicle trajectories provide data that is superior to conventional sensor data, as observation is independent of spatial restrictions and unaffected by queue buildup and discharge. Thus intersection approach performance is no longer estimated but measured, at multiple levels from isolated intersections to arterials and networks. Vehicle trajectory data superimposed with signal phase indication duration create distinguishable time-space-signal (TSS) signatures that provide insightful diagnostic capabilities to uncover possible reasons of inferior performance of traffic signal systems. This study provides a diagnostic tool aimed at determining traffic control operational deficiencies, as well as the extent to which deployed signal timing plans successfully respond to prevailing traffic conditions. A practical and straightforward approach was designed to visualize and analyze individual phases’ green time utilization in a significant number of operational scenarios. The practicality of the proposed method is reflected in reducing the time and effort required by the existing signal design/retiming practice, since trajectory-signal signatures distinguish between incidents and retiming opportunities caused by changing traffic conditions. The conceptual framework for a high-resolution data analysis platform formulated in this study identifies the causes of intersection performance deterioration, by defining a set of visual and quantitative operational success indicators.

William Alexander, University of Texas, Austin
Alexander Hainen, University of Alabama
Xiaoyan Hong, University of Alabama

Vehicle-to-Infrastructure communications offer great opportunities for developing applications for Intelligent Transportation Systems with real-time vehicle motion information. In this paper, we propose to leverage this technology in a system that reduces delay for a priority vehicle by adjusting an isolated, actuated traffic signal controller’s timing mechanism in a non-invasive manner. We developed an algorithm and architecture that monitors connected vehicle location data and recent intersection performance, calculates a minimum-disruption timing plan which provides a green light to the priority vehicle at its arrival, if possible, then enforces the plan until the vehicle arrives. The algorithm and software architecture were tested in simulation, with results showing significant reduction in priority vehicle delay with little impact on traffic at the intersection. The work demonstrates the power of Vehicle-to-Infrastructure communication in enhancing transportation operation. The results show that, by temporarily enforcing a timing plan and modifying it to accommodate a priority vehicle’s arrival, the delay the vehicle encounters at an intersection can be reduced without adding significant delay for other vehicles crossing said intersection. The simulation results we present indicate that this system could be deployed at an isolated intersection to dramatically reduce priority vehicle delay without significant detriment to the remaining vehicles utilizing the intersection. By extending this system with vehicle occupancy data, decision-making can be made autonomously as to whether, when, and how to facilitate a priority vehicle while minimizing impact to the rest of the operations at an intersection.

Jian Zhang, Southeast University
Shuyang Dong, Southeast University
Han Wang, Southeast University
Rui Li, Hohai University
Bin Ran, Southeast University

This paper proposed a CVIS-based Eco-driving control model for divisible electric platoons at signalized intersection. Platoon releasing plan and signal control scheme are obtained through platoon splitting and speed guidance. The method builds a multi-objective optimization model with the objective functions aiming at minimizing platoon passing time and energy consumption and the optimal solution is obtained by genetic algorithm. According to the optimal scheme the platoons are split up and speed guidance is conducted. Combining with the phase plan, the platoons are released at the signalized intersection. The simulation of proposed Eco-driving platoon control model (EPCM) and actuated platoon control model (APCM) were conducted with Matlab and Simulation of Urban Mobility (SUMO) invoked by Python. Simulation results shows the average waiting time of vehicles decreased by 8.97%, the maximum queuing time of vehicles in different directions decreased from 38.1% to 54.3%, and the total energy consumption decreased by 21.3%, which indicate that the EPCM can effectively reduce the energy consumption of electric platoons while through the intersection as well as ensuring the passing efficiency and reducing the average waiting time at the stop line.

Jiarong Yao, Tongji University
Chaopeng Tan, Tongji University
Keshuang Tang, Tongji University

Traditional arterial coordination control models adopt aggregated demand from fixed detectors or manual counts as main input, which lacks a self-feedback mechanism between the input and the optimization objective. With the population of connected vehicles and intelligent mobility, high-resolution trajectory data provide new possibilities for signal control evaluation and optimization. Therefore, the objective of this study is to propose a new arterial coordination control model for two-way arterial progression solely using sampled trajectories. Sampled trajectories are first used to extract prior arrival information and queuing states. Then, vehicle arrival time at the stop-line at each intersection is estimated as the function of signal timing parameters and prior arrival rates based on the shockwave theory, considering different sampled trajectory statuses. The delay of each sample vehicle is thus obtained as the difference between the actual arrival time and the projected arrival time. Sum of the delays of all the sampled trajectories at each intersection is selected to be the optimization objective, and the optimization problem is solved through a multi-sub-swarm Particle Swarm Optimization (PSO) algorithm. A simulation model is built to test the performance of the proposed model compared with the Synchro model and the MULTIBAND model under various demand scenarios. Results show that the performance of the proposed method is about 21% better than the Synchro model and 14% worse than the MULTIBAND model. This work demonstrates that the optimization of fixed-time arterial coordination control solely using sample trajectories is feasible, and a satisfactory performance can be expected as well.

Yu Mei, Hong Kong Polytechnic University
Fuliang Li, Didi Chuxing LLC
Jianfeng Zheng, Didi Chuxing LLC
Yu Han, Didi Chuxing LLC
Henry Liu, University of Michigan, Ann Arbor
Cheng Gong, Didi Chuxing

SCATS (Sydney Coordinated Adaptive Traffic System) imbeds two green split optimization strategies: discrete split selection (DSS) and incremental split selection (ISS). In China over 6500 intersections are under the control of SCATS and most of them have applied DSS strategy. Its main reason is that a great number of detectors working for SCATS are in poor conditions and ISS is sensitive to errors of detectors. The performance of DSS strategy is dominated by a preset signal plan table, which is currently designed manually by SCATS practitioners. However, a good design is costly and highly depends on the experience of designers, which makes it intractable for frequently plan adjustments to accommodate time-variant traffic demand. This paper proposed a method to optimize the signal plan table for SCATS-controlled intersections with the use of vehicle trajectory data. By designing quality split plan tables, it can make the SCATS adaptively adjusts green split for the traffic even through some detectors were reporting incorrect data. The proposed method is tested in real world at a SCATS intersection in Guangzhou, China. The results show the method is able to improve the performance of the study intersection, especially under the off-peak traffic conditions.

Tao Fu, Tongji University
Wanjing Ma, Tongji University
Ling Wang, Tongji University
Fan Zhang, Research Institute of Highway, Ministry of Transport

The efficiency evaluation of signal control intersection is the basis of signal optimization, and green time utilization is one of the most important evaluation indicators. In current studies, the most commom input of evaluation of the green time utilization for pre-timed signals is traffic volume data from fixed detectors. Fixed detectors might not be accurate due to detector damage and might not be available because of low coverage in some cities. On the other hand, with the development of Global Position System technology, the possibility of applying trajectory data to evaluate the efficiency of pre-timed signals has been continually gaining interest. This paper proposed a method for evaluating the utilization of green time for pre-timed signals using trajectory data. Based on the stop-wave and dissipation-wave information contained in the trajectory data, three indicators, Overall Green Time Utilization Rate, Insufficient Occupancy Rate, and Skewness of Occupancy, were proposed to comprehensively evaluate the utilization of green time. A field case study was used to show the effectiveness of this methodology, and a sensitivity analysis was carried out by VISSIM simulation to illustrate the applicability. The results showed that when the penetration rate reaches 10% with a two hours database or when the total sampling time reaches four hours under the penetration rate of 5%, the mean absolute percent error value of the overall green time utilization rate will be less than 10% and the performance will be relatively stable.

Matthew Muresan, University of Waterloo
Liping Fu, University of Waterloo
Ming Zhong, Wuhan University of Technology

This paper discusses a method to adjust traffic signal timings in real-time using travel time data obtained from Bluetooth and WiFi sensors. The proposed method attempts to leverage the opportunity of increasing availability of “Big Data” for improving the prediction of traffic characteristics. This type of data is widely available but has not seen use beyond traffic monitoring. A peudo-adaptive signal updating is proposed on the basis of an analytical approach from Highway Capacity Manual (HCM), enabling the best use of the available data while keeping the computational time low. A simulation platform (VISSIM) with scenarios covering a wide variety of penetration rates and traffic configurations is used for testing and development. Bluetooth and WiFi detections based on typical technical parameters are simulated in an external module developed to connect to the simulator. A sensitivity analysis is conducted to assess the effect of some key parameters such as the frequency of signal adjustment, the size of the adjustment, and the market penetration. Performance data is also extracted from multiple simulation runs to contrast the benefits of the proposed system compared to the traditional methods. The results of the sensitivity analysis show that the system can perform well in situations with penetration rates as low as 10% and adjustment intervals as short as 5 minutes. Travel time delays are also reduced when compared to semi-actuated control.

Lijuan Wan, Tongji University
Wanjing Ma, Tongji University
Chunhu Yu, Tongji University
Ling Wang, Tongji University

The time-of-day (TOD) mode is the most widely used strategy for traffic signal control with fluctuating flows. Most studies determine TOD breakpoints based on traffic volumes collected by infrastructure-based detectors (e.g., loop detectors). However, such infrastructure-based detectors have low coverage and high maintenance cost. With the deployment of probe vehicles, vehicle trajectory data become available providing a new data source for signal control. This paper proposes an approach to identify TOD breakpoints at an isolated intersection based on the trajectory data of probe vehicles, instead of conventional traffic volumes, with under-saturated traffic condition. It is shown that the speeds of queueing shockwaves capture the characteristics of traffic volumes according to the queueing shockwave theory. The data of multiple sampling days are aggregated to compensate for the limitations of low market penetration rates and long sampling intervals. Queue joining vehicles are then identified to obtain the speed of queueing shockwaves. The bisecting K-means algorithm is applied to cluster periods, which are characterized by queueing shockwave speeds, to identify TOD breakpoints. The numerical studies validate that the speeds of queueing shockwaves capture the trend of traffic volumes. The clustering algorithm identifies the same TOD breakpoints for queueing shockwave speeds and traffic volumes. As long as the number of sampling days is large enough, the proposed method can handle low penetration rates (e.g., 2%) and long sampling intervals (e.g., 20 s), and thus achieve comparable performance as the ideal condition with high penetration rates (e.g., 100%) and short sampling intervals (e.g., 1 s).

Nemanja Dobrota, Florida Atlantic University
Aleksandar Stevanovic, Florida Atlantic University
Nikola Mitrovic, Florida Atlantic University
Sharmin-E-Shams Chowdhury, Florida Atlantic University

Signal retiming or optimization process has not been changed much over the last few decades. Very often, more is achieved through the process of “fine-tuning” than by signal optimization itself. New trend suggest use of the high-resolution data but many improvements could be made if the traditional cyclical signal data are utilized in better way. This study covers such an attempt where large data sets, high-fidelity simulation models, and powerful stochastic optimizations are utilized to develop more robust signal timing plans. The study covers a 28-intersection urban corridor in Southeastern Florida. Historical data are used to identify Day-Of-Year (DOY) representative volumes for major seasons for a peak hour. Microsimulation models are developed and calibrated with high precision to match the field signal operations. Then signal timings are developed, by using traditional and stochastic optimization tools, for those DOY plans and tested in microsimulation environment. The findings reveal shortcomings of the traditional tools which cannot model intricacies of the field operations. Results from stochastic optimizations reduced average delay between 5% and 26%. Stops were decreased with a smaller magnitude (1-18%) but the final solution certainly did not transfer delay on the other neighboring roads since the latent delay (representing traffic from surrounding network) also decreased by 10-49%. In general the use of this method has shown a great potential and the next step should be field testing and validation.

Hao Wu, Tongji University
Jiarong Yao, Tongji University
Lei Liu, Tongji University
Yumin Cao, Tongji University
Keshuang Tang, Tongji University

Left-turn queue spillback occurs when the queue of left-turning traffic or adjacent through traffic exceeds storage space of the left-turn lane. It leads to discharge flow breakdown and ineffective green time, and thus significantly deteriorates operational efficiency of signalized intersections. License Plate Recognition (LPR) systems widely implemented at signalized intersections in China can record individual vehicles’ departure time at the stop-line of each approach lane, which offers an ideal data source for the identification of queue spillback. Thus, this paper presents a left-turn queue spillback identification method based on LPR data. In this study, ten types of left-turn queue spillbacks are firstly defined according to queuing status of the left-turn lane and the adjacent through lanes. Then, a lane-based queue length estimation algorithm incorporating the Critical Point Analysis (CPA) is proposed to estimate the queue lengths at the left-turn lane and the adjacent through lanes. Finally, a rule-based identification algorithm is developed to specify queue spillback types, according to the left-turn phasing, the estimated queue lengths, and the length of storage space. The proposed method is evaluated using traffic simulation under various scenarios, in which five left-turn phasing schemes are tested and compared. Results indicate that the successful identification rate of the proposed method for all the left-turn phasing schemes is nearly 90% on average, and achieves the highest 96% for the lagging and protected-only left-turn phase. This work can further be applied to signal control optimization at intersections, particular for the oversaturated traffic conditions.

Zhe Zheng, Tongji University
Wanjing Ma, Tongji University
Jing Zhao, University of Shanghai for Science and Technology

As one type of the unconventional signalized intersections, the tandem intersection could improve the capacity of the interaction with the optimal control of the pre-signal and the main-signal. In order to enhance the performance of the tandem intersection, the delay model and the capacity model was established based on the surveryed intersection. Considering about the stochastic variance of the volume at the pre-signal and the stochastic variance of saturation flow rate of lanes in the sorting area, a robust optimal model of signal timing was proposed. The objects of the model are twofold: efficiency and robustness, including standard deviation, the half standard deviation, and the most negative value of the delay of the intersection. The accuracy of the proposed robust optimal model was verified and compared with the existing models. A case study was used to test the effect of the proposed model. The results show that the control effect of the proposed model is better than the traditional model by 27.84% in delay and 22.92% in capacity. Meanwhile, the delay of the tested intersection for stochastic variance of the volume and saturation flow rate could be fully improved with a 31.22% reduction.

Jiaming Wu, Southeast University
Pan Liu, Southeast University
Yang Zhou, University of Wisconsin, Madison

The present paper conducted a stochastic analysis of the operational performance of the CLL design considering the influence of the upstream signalized intersection. The arrival distribution was generated using the platoon dispersion model. A stationary condition in which the number of discharged left-turning vehicles in both the contraflow lane and the conventional lane would be constant in any stationary cycles was defined. We proved that the CLL system will always reach the stationary condition after a few cycles if the arrival distribution is fixed. In stationary cycles, the CLL design consistently generates either recurrent and constant residual queues or no queues, depending on the arrival distributions of left-turning vehicles. Considering the stationary condition, analytical models were developed for estimating the capacity and delay for left-turns at signalized intersections with the CLL design. The results showed that both the arrival pattern and the length of the contraflow lane could significantly influence the operational performance of the CLL design. The residual queues in the stationary condition might significantly increase control delay, indicating that left-turning vehicles could experience longer delay if the contraflow lane is not properly designed. An optimization method was then proposed for minimizing the control delay by optimizing the length of contraflow lanes and the offset between intersections considering fixed left-turn demand. The research results can be directly used by traffic engineers to optimize the CLL design, and to estimate the operational performance of the signalized intersections with the CLL design.

Marija Ostojic, Northwestern University
Archak Mittal, Ford Motor Company
Hani Mahmassani, Northwestern University
David Hale, Leidos, Inc.

Since the 2000’s, researchers have been designing and examining the potential of visualizing high-resolution signal and vehicle trajectory data for the purpose of gaining better insight into traffic control system performance, along with other representative measures of effectiveness. Along with technological advancements, obtaining finer granularity data became achievable. Visualizing relevant signal performance data in an easy-to-understand format is crucial when identifying problem areas and/or causes of problems. Furthermore, visualizing of now-obtainable information needs to be accompanied by a tangible and robust performance metric. To analyze this high-quality information when assessing the quality of signal timing settings, new performance metrics are required since traditional no longer suffice the definition of state-representativeness. The proposed rate of utilization of available green time and space incorporates multiple aspects of signal performance assessment: total delay, progression speed, quality of progression, as well as the impact of oversaturation. It can be applied at an individual vehicle level, measuring how well the vehicle is utilizing time and space under specific signal control operational conditions, or any other more aggregate level. The concept utilizes vehicle-trajectory and signal phase status and duration data to ascertain the responsiveness of trajectory-based measures when different control strategies are applied. To develop and demonstrate the concept, the study uses simulation data in a format that corresponds to high resolution data - signal status and vehicle positions at a one tenth of a-second frequency.

Tao Fu, Tongji University
Chunhu Yu, Tongji University
Li Zou, DiDi Smart Transportation
Weili Sun, Didi Chuxing
Cheng Gong, Didi Chuxing
Wanjing Ma, Tongji University

Pre-timed signal control is widely applied to improve mobility and safety at isolated intersections. Traffic volumes, which are collected by fixed detectors, are commonly used as the input to the optimization of signals. However, such fixed detectors have low coverage and high maintenance cost because of the damage in practice, especially, in developing countries. The deployment of probe vehicles equipped with Global Position System (GPS) makes vehicle trajectory data available, which provides a new data source for signal timing. This paper proposes an approach to optimize signals (i.e., phase sequences, green splits, and cycle lengths) based on vehicle trajectory data. No estimation of traffic volumes or assumptions of vehicle arrival patterns are made to handle low penetration rates of probe vehicles. The impacts of signal timing on vehicle trajectories are analyzed. Nonlinear programming models are formulated with the focus on the overall benefits of vehicles in all approaches (OPO model) and the worst average benefits of vehicles in different approaches (APO model). The nonlinear models are solved by solving a series of mixed-integer-linear-programming models. The proposed models are validated by numerical studies. It outperforms the Webster model significantly, especially with heavy traffic. The OPO model is preferred if the penetration rates are similar in different approach. Otherwise, the APO performs better.

Jonathan Waddell, Wayne State University
Stephen Remias, Wayne State University
Jenna Kirsch, Wayne State University

Performance measures offer an essential management tool for transportation engineers to make decisions along signalized corridors. Coordination is a critical component of signal timing, which enables the progression of vehicles through a corridor, reducing travel time, delay, and the stopping of vehicles. Current signal performance strategies assess coordination of a corridor using intersection level metrics while relying on expensive infrastructure to measure vehicle arrivals. Recent crowdsourced data collection strategies have allowed for the ubiquitous collection of individual vehicle waypoints. These trajectories can be used to replicate existing signal performance measures and improve upon current practices. This paper uses trajectory data from numerous corridors around the State of Michigan to illustrate the merit and versatility of floating car based performance measures. Concepts including data validation, data aggregation, performance thresholds, outcome assessment, and reliability are replicated using trajectory data. The paper concludes by summarizing historically unattainable corridor level performance and reliability. Findings from this paper show current Automated Traffic Signal Performance Measures (ATSPM) can be replicated using low penetration rate vehicle trajectory data, such as percent arrival on green and delay quantification. The metrics derived from trajectories are then compared on two Michigan corridors to those derived from ATSPMs. Also the reliability based performance metric, Level of Travel Time Reliability (LOTTR), can be improved using trajectories of vehicles known to travel the complete corridor instead of aggregating segmented probe vehicle data. The LOTTR metric is compared using eight corridors from Michigan with trajectory data and segment based crowd sourced probe vehicle data.

Lei Liu, Tongji University
Chaopeng Tan, Tongji University
Jiarong Yao, Tongji University
Yumin Cao, Tongji University
Keshuang Tang, Tongji University

Queue length is one of the most important performance measures for signalized intersections. With recent advancements in connected vehicles and intelligent mobility technology, utilizing vehicle trajectory data to estimate queue length has received considerable attentions. However, most of the existing methods are based on some assumptions, such as known arrival patterns, and/or high penetration rates. Besides, the existing models would probably be unstable or invalid under sparse trajectory environment. Hence, License Plate Recognition (LPR) data is introduced in this study to fuse with the vehicle trajectory data, and then a lane-based queue length estimation method is proposed. Firstly, based on the Kernel Density Estimation method, the probability density distribution of discharge headway and the stop-line crossing time for the queued and non-queued vehicles are obtained by using historical trajectory data as well as LPR data. Then, the Bayesian method is used to derive queue length with the largest possibility. Finally, the performance of the proposed method is evaluated through simulation and empirical cases. Results show that the proposed method could obtain accurate estimation of queue length for most conditions, and achieve acceptable accuracy even under a relatively low penetration rate (3%). Under the extreme circumstance that no probe vehicles are captured during a cycle, a reliable estimation could also be produced by the proposed method.

Michael Dunn, University of Texas, Austin
Heidi Westerfield Ross, University of Texas, Austin
Carolina Baumanis, University of Texas, Austin
Jared Wall, The City of Austin Transportation Department
Jonathan Lammert, The City of Austin Transportation Department
Jen Duthie, The City of Austin Transportation Department
Natalia Ruiz-Juri, University of Texas, Austin
Randy Machemehl, University of Texas, Austin

Signal retiming is one of the chief responsibilities of municipal transportation agencies, and is an important means for reducing congestion and improving transportation quality and reliability. Many agencies conduct signal retiming and adjustment in a schedule-based manner. However, leveraging a data-driven, need-based approach to signal retiming to prioritize operations could better optimize use of agency resources. Additionally, the growing availability of probe vehicle data has made it an increasingly popular tool for use in roadway performance measurement. This paper presents a methodology for utilizing segment-level probe-based speed data to rank the performance of traffic signal corridors for retiming purposes. This methodology is then demonstrated in an analysis of 79 traffic signal corridors maintained by the City of Austin, Texas. The analysis considers 15-minute speed records for all weekdays in September 2016 and September 2017 to compute metrics and rank corridors based on their relative performance across time periods. The results show that the ranking methodology compares corridors equitably despite differences in road length, functional class, and traffic signal density. Additionally, results indicate that the corridors prioritized by the ranking methodology represent a much greater potential for improving travel time than the corridors selected under the schedule-based approach.

Zelong Du, DiDi Smart Transportation
Xintao Yan, DiDi Smart Transportation
Jinqin Zhu, Didi Chuxing LLC
Weili Sun, Didi Chuxing

Signal phase and timing (SPaT) information is critical for many in-vehicle applications. However, it is challenging and time-consuming to acquire city-wide SPaT information from local traffic management agencies directly. A significant limitation of existing SPaT information estimation methods in the literature is that they can only be applied to a specified Time-of-Day (TOD) period. In the real-world, however, different TOD timing plans are used to accommodate fluctuations in traffic demands. In this paper, we propose a novel method for traffic light parameter estimation based on Floating Car Data (FCD), which features recognizing TOD breakpoints and thus can be applied to intersections with multi-TOD timing plans. Also, the good estimation of TOD breakpoints leads to more data availability for estimation of other parameters. The proposed method is tested with real-world data collected from DiDi on-line hailing platform in China. The filed test results show promising accuracy. The absolute error of green duration is within 3 seconds in daytime and the estimation error of TOD breakpoints is within 15 minutes.