Quantitative Evaluation of Advanced Traffic Management Systems using Multi-Criteria Decision Analysis
Mohammad Razaur Rahman Shaon (firstname.lastname@example.org), University of ConnecticutShow Abstract
Xiaofeng Li, University of Arizona
Yao-Jan Wu, University of Arizona
Simon Ramos, City of Phoenix
With the technological advancements in recent years, a series of Intelligent Transportation System (ITS) products have now become available to the transportation agencies to collect data and manage traffic conditions on the roadway network. Among ITS products, Advanced Traffic Management System (ATMS) has been effectively serving as the central nervous system of a Traffic Management Center. ATMS serves as an integrated application for a wide variety of purposes ranging from data collection to implementing traffic management strategies. Due to commercial popularity, a series of ATMS products are now available to the transportation agencies and there is no consensus on selecting the best-suited product based on tailored requirements. Making a decision for a decision-critical item such as ATMS products on qualitative evidence can add risk to the decision-makers to justify their decision of choice. In this study, a Multi-criteria Decision Analysis framework was proposed for quantitative evaluation of ATMS alternatives that can consider multiple and conflicting decision-making criteria using a real-world example. Moreover, the proposed framework was evaluated for different scenarios related to different applications of ATMS products to provide flexibility to the user in evaluating the ATMS alternatives. Results indicated that the proposed method can be considered as a viable alternative in contrast to a qualitative evidence-based decision-making strategy to minimize the risk associated with the decision-makers. Using the proposed quantitative framework, decision-makers can examine the weights of different criteria under consideration and evaluate multiple ATMS alternatives. The proposed framework can be easily applied to other ITS technology selection processes.
A Study on Feasibility of Vehicle to Infrastructure Applications Supported Through Advanced Wireless Communications
Anjan Rayamajhi, Leidos, Inc.Show Abstract
Animesh Balse, Leidos, Inc.
Sudhakar Nallamothu, Leidos, Inc.
Hyungjun Park, Federal Highway Administration (FHWA)
Connected and automated vehicle (CAV) technology has the potential to improve transportation systems. It has been shown to bring benefits in transportation mobility, safety, and the environment. The signal phase and timing (SPaT) message is one of the fundamental and critical CAV messages since it enables the connectivity between vehicles and the infrastructure. Because SPaT messages can be transmitted using different wireless communication technologies, it is necessary to study latency and coverage of SPaT messages in such cases. This study investigates performances of SPaT transmitted using two popular communication technologies in CAV – dedicated short range communications (DSRC) and cellular 3GPP 4G/LTE (the 3 rd generation partnership project/4 th generation long-term evolution). To provide a robust evaluation, SPaT data transmitted by DSRC and cellular were collected in the field at various intersections in Northern Virginia and compared in terms of latency and distance coverage. The results showed that the latency experienced by SPaT messages over cellular networks is well below 100 milliseconds required by most infrastructure applications, implying that cellular communications may be used for vehicle to infrastructure (V2I) based safety and mobility applications. The feasibility of several CAV applications was investigated based on the network performance observed in this study. Specifically, several safety, mobility and environment applications were tested to see if DSRC and cellular could satisfy their requirements in terms of network performance. The study also provides conclusive remarks on whether the wireless communication technologies are capable of supporting such applications.
Automatic Traffic Queue-End Identification Using Location-Based Waze User Reports
Yuandong liu (email@example.com), Oak Ridge National LaboratoryShow Abstract
Zhihua Zhang, University of Tennessee, Knoxville
Lee Han, University of Tennessee
Candace Brakewood, University of Tennessee, Knoxville
Traffic queues, especially queues caused by non-recurrent events such as incidents, are unexpected to high-speed drivers approaching the end of queue (EOQ) and become safety concerns. Though the topic has been extensively studied, the identification of EOQ has been limited by the spatial-temporal resolution of traditional data sources. This study explores the potential of location-based crowdsourced data, specifically Waze user reports. It presents a dynamic clustering algorithm that can group the location-based reports in real-time and identify the spatial-temporal extent of congestion as well as the end of queue. The algorithm is a spatial-temporal extension of DBSCAN (density-based spatial clustering of applications with noise) algorithm for real-time streaming data with an adaptive threshold selection procedure. The proposed method was tested with 34 traffic congestion cases in the Knoxville,Tennessee area. It is demonstrated that the algorithm can effectively detect spatial-temporal extent of congestion based on Waze report clusters and identify end of queue in real-time. The Waze report-based detection are compared to the detection based on road side sensor data. The results are promising: The EOQ identification time of Waze is similar to the EOQ detection time of traffic sensor data, with only 1.1 minutes difference on average. In addition, Waze generates 1.9 EOQ detection points every mile, compared to 1.8 detection points generated by traffic sensor data, suggesting the two data sources are comparable in terms of reporting frequency. The results indicate that Waze is a valuable complementary source for end of queue detection where no traffic sensors are installed.
Evaluation of Wrong-Way Driving (WWD) Detection Proof of Concept
Alvin Stamp, Colorado Department of TransportationShow Abstract
Susi Marlina, Colorado Department of Transportation
Bruce Janson, University of Colorado, Denver
Vijay Sabawat, Felsburg Holt and Ullevig
Wrong-way driving (WWD) incidents often result in one of the most serious types of crashes to occur on our interstate highways, often resulting in fatalities. According to the Federal Highway Administration (FHWA), 300 to 400 fatalities are caused annually by wrong-way driving in the United States. Colorado Department of Transportation (CDOT) Region 1 (R1) conducted a pilot study to explore different types of available WWD detection technologies for interstate off-ramps to evaluate which system was most accurate in detecting wrong-way vehicles in a variety of different weather conditions. The pilot study evaluated multiple different wrong-way detection systems: one video camera, one radar, and two different thermal cameras at Interstate 70 and Ward Road. The WWD detection system vendors were responsible to install, maintain, and operate the required communications systems and to send an alert via email to the project team whenever a WWD vehicle is detected. All system vendors opted for cellular communications to send the alerts. For this study, no additional infrastructure such as a LED beacon was installed to alert the wrong-way driver. Two systems were found to have similarly high odds ratios in detecting wrong-way vehicles. These two systems were found to be superior in both daytime and nighttime conditions, which did not impact the accuracy of these two detection systems.
Sequential Optimization of an Emergency Response Vehicle’s Intra-link Movement in a Partially Connected Vehicle Environment
Gaby Joe Hannoun (firstname.lastname@example.org), New York University, Abu DhabiShow Abstract
Pamela Murray-Tuite, Clemson University
Kevin Heaslip, Virginia Polytechnic Institute and State University (Virginia Tech)
Thidapat Chantem, Virginia Polytechnic Institute and State University (Virginia Tech)
This paper introduces a semi-automated system that facilitates Emergency Response Vehicle (ERV) movement through a transportation link by providing instructions to downstream non-ERVs. The proposed system adapts to information from non-ERVs that are nearby and downstream of the ERV. As the ERV passes stopped non-ERVs, new non-ERVs are considered. The proposed system sequentially executes integer linear programs (ILPs) on transportation link segments with information transferred between optimizations to ensure ERV movement continuity. This paper extends a previously developed mathematical program that was limited to a single short segment. The new approach limits runtime overhead without sacrificing effectiveness and is more suitable to dynamic systems. It also accommodates partial market penetration of connected vehicles using a heuristic reservation approach, making the proposed system beneficial in the short-term future. The proposed system can also assign the ERV to a specific lateral position at the end of the link, a useful capability when next entering an intersection. Experiments were conducted to develop recommendations to reduce computation times without compromising efficiency. When compared to the current practice of moving to the nearest edge, the system reduces ERV travel time an average of 3.26 seconds per 0.1 mile and decreases vehicle interactions.
Intersection Braking Advisor: A Connected Vehicle-Infrastructure Application
David LeBlanc, University of Michigan, Transportation Research InstituteShow Abstract
Scott Bogard, University of Michigan, Ann Arbor
Kathiravan Natarajan, Honda R&D Americas, Inc.
Sue Bai, Honda Research Institute Europe GmbH
A new connected vehicle application -- Intersection Braking Advisor– is proposed to provide drivers with in-vehicle cues when the application predicts that braking at an upcoming equipped intersection is likely to provide the safest outcome. Intersection Braking Advisor uses signal phase and timing broadcasts and genuine or proxy basic safety messages from the vehicle ahead, to predict the likely outcome of not slowing, as well as the necessary braking to stop at the stop bar. Estimates of preceding vehicle location trajectories and also used. The main effect of the application is to create the equivalent of an in-vehicle, earlier and longer yellow phase, as well as providing a possibly more accurate and reliable prediction of whether slower traffic ahead should affect the decision to brake or to continue through the intersection. This paper presents algorithms for the two cases of being unimpeded or impeded by slower traffic ahead on the approach to the signal. Simulations are conducted to show the likely outcome of such approaches. The next steps would include simulating the application’s performance across a variety of intersections and real-world timing plans and to incorporate naturalistic driving data distributions of driver braking at intersections to refine the algorithms and further focus the potential benefits. Human factors studies would be recommended before actual deployment testing proceeds.
Traffic Density Estimation via a Particle Filter using Connected Vehicle Data
Mohammad Aljamal, Virginia Polytechnic Institute and State University (Virginia Tech)Show Abstract
Hossam Abdelghaffar, Virginia Polytechnic Institute and State University (Virginia Tech)
Hesham Rakha (email@example.com), Virginia Polytechnic Institute and State University (Virginia Tech)
The paper presents a nonlinear filtering approach to estimate the traffic stream density on signalized approaches based solely on connected vehicle (CV) data. Specifically, a particle filter (PF) is developed to produce reliable traffic density estimates using CV travel time measurements. Traffic flow continuity is used to derive the state equation, whereas the measurement equation is derived from the hydrodynamic traffic flow relationship. Subsequently, the PF filtering approach is compared to simple linear estimation approaches; namely, a Kalman filter (KF) and an adaptive KF (AKF). Simulated data are used to evaluate the performance of the three estimation techniques on a signalized approach experiencing under- and over-saturated conditions. Results demonstrate that the three techniques produce accurate estimates, with the KF, surprisingly, being the most accurate of the three techniques. A sensitivity of the estimation techniques to various factors including the CV level of market penetration, the initial conditions, and the number of particles in the PF is also presented. As would be expected, the study demonstrates that the accuracy of the PF estimation increases as the number of particles increases. Furthermore, the accuracy of the density estimate increases as the level of CV market penetration increases. The results indicate that the KF is least sensitive to the initial vehicle count estimate, while the PF is most sensitive to the initial condition. In conclusion, the study demonstrates that a simple linear estimation approach is best suited for the proposed application.
Vehicle Trajectory Optimization at Roundabouts with a Mixed-fleet of Automated and Human-driven Vehicles
Rasool Mohebifard, North Carolina State UniversityShow Abstract
Ali Hajbabaie (firstname.lastname@example.org), North Carolina State University
This paper presents a methodology to control the trajectory of connected automated vehicles (CAVs) at roundabouts operating with a mixed fleet of CAVs and human-driven vehicles (HVs). We formulate a non-convex optimization program in a two-dimensional space for this purpose. A model predictive control-based solution technique is developed to optimize the trajectories of CAVs at discretized time steps based on the estimated driving behavior of HVs, while the actual behavior of HVs is controlled by a microscopic traffic simulator. At each time step, the location and speed of vehicles are collected and a decomposition-based methodology optimizes the trajectories for a few time steps ahead of the system time. The optimization methodology has convexification, alternating direction method of multipliers, and cutting plane decompositions to tackle the complexities of the problem. We implemented the solution technique in a case study roundabout with different demand levels and penetration rates of CAVs. We also compared the results with a simulation-based technique to evaluate the effects of CAVs on traffic operations. The results showed that traffic operations were consistently improved with the presence of CAVs ranging from 20% to 100% penetration rates. Noticeable improvements in the safety measures were also observed at the CAV penetration rates greater than or equal to 60%.
Multinomial Classification Method for Short-Term Spatial and Temporal Prediction of Freeway Crashes: Case Study of Interstate 270 in Missouri
Osama Mohammed, Saint Louis UniversityShow Abstract
Jalil Kianfar (email@example.com), Saint Louis University
Yadong Wang, Southern Illinois University, Edwardsville
Roadway incidents are a major source of non-recurring congestion in freeway systems. Transportation agencies implement traffic incident management systems to reduce the impacts of incidents. This paper proposes a machine-learning approach to predict locations and types of incidents on a roadway system during short intervals. This information will allow transportation agencies to proactively allocate their resources in anticipation of roadway incidents. Multinomial ensemble learning-based classification models were developed to predict the number of incidents, their type, and roadway sections where crashes will occur within a one-hour prediction period. Random forest and gradient boosting machines in conjunction with a binary relevance transformation approach and label power set transformation approach were used to predict crashes for a segment of Interstate 270 in St. Louis, Missouri. The case study results indicated that models developed based on the random forest method outperform the model developed based on gradient boosting machines method. The precision, recall, and F1 score of classifications obtained on the basis of the label power set transformation approach were slightly better than the precision, recall, and F1 score of classifications obtained from the binary relevance approach.
A Pattern-Based Automatic Incident Detection Method for a Vehicle-to-Infrastructure Communication Environment
Kun Zhang, Saint Louis UniversityShow Abstract
Jalil Kianfar (firstname.lastname@example.org), Saint Louis University
Kamran Moghaddam, Clayton State University
Transportation agencies continuously and consistently work to improve the processes and systems for mitigating the impacts of roadway incidents. Such efforts include utilizing emerging technologies to reduce detection and response time to roadway incidents. Vehicle-to-Infrastructure (V2I) communication is an emerging transportation technology that enables communication between a vehicle and the infrastructure. This paper proposed an algorithm that utilized V2I probe data to automatically detect roadway incidents. A simulation test-bed was developed for a segment of Interstate 64 in St. Louis, Missouri to evaluate the performance of the V2I-based automatic incident detection algorithm. The proposed algorithm was assessed during peak and off-peak periods with various incident durations, under several market penetration rates for V2I technology, and with different spatial resolutions for incident detection. The performance of the proposed algorithm was assessed on the basis of detection rate, time to detect, detection accuracy, and false alarm rate. The performance measures obtained for the V2I-based automatic incident detection algorithm were compared with California #7 algorithm performance measures. California #7 algorithm is a traditional automatic incident detection algorithm that utilizes sensors such as inductive loop detectors to identify roadway events. The California #7 algorithm was implemented in Interstate 64 simulation test-bed. The case study results indicated that the proposed algorithm outperformed the California #7 algorithm. The detection rate for the proposed V2I-based incident detection algorithm was 100% in market penetrations of 50%, 80%, and 100%. However, the California #7 algorithm detection rate was 71%.
Exploring Dynamic Mode Decomposition for Robust System Identification: Applications to Adaptive Signalised Intersections
Kazi Redwan Shabab, University of Central FloridaShow Abstract
Shakib Mustavee, University of Central Florida
Shaurya Agarwal, University of Central Florida
Mohamed H. Zaki (email@example.com), University of Central Florida
This paper introduces a novel data-driven approach based on recent developments in dynamic mode decomposition (DMD) for system identification of queue formation at signalized intersections. Traffic systems such as vehicular flow and queue formation on signalized intersections have complex nonlinear dynamics, making system identification and controller design tasks challenging. We explore dynamic mode decomposition with control (DMDc) and Hankel DMD with control (HDMDc) for system identification and to obtain locally linear dynamics. We analyze several key aspects and provide insights into the DMD application. For instance, we analyze the impact of delay embedding and selection of the numbers of training snapshots in system identification results using DMD based algorithms. To demonstrate the application of the obtained linearized dynamics, we perform prediction of the queue lengths at the intersection; and compare the results with the state-of-the-art long short term memory (LSTM) method. The case study involves the morning peak vehicle movements and queue lengths at two signalized intersection in Orlando area. It is observed that DMD based algorithms are able to capture the complex dynamics with a linear approximation to a reasonable extent. It is expected that this linear yet accurate system identification approach will be beneficial in addressing many challenges in smart mobility, including designing controllers for an adaptive traffic signal.
Video-based Network-wide Surrogate Safety Analysis to Support a Proactive Network Screening Using Connected Cameras: Case Study in the City of Bellevue (WA) United States
Lana Samara, Transoft Solutions, Inc.Show Abstract
Paul St-Aubin, Transoft Solutions, Inc.
Franz Loewenherz, City of Bellevue
Noah Budnick, Together For Safer Roads
Luis Miranda-Moreno, McGill University
Surrogate road safety approaches, as part of road improvement programs, have gained traction in recent years. Thanks to emerging technologies such as computer-vision and cloud-computing, surrogate methods allow for proactive scanning and detection of safety issues and address them before collisions and injuries occur. The objective of this paper is to propose an automated and continuous monitoring approach for road network screening using connected video cameras and a cloud-based computing analytics platform for large-scale video processing. Using the wide network of traffic cameras from cities, the proposed approach aims to leverage video footage to extract critical data road network screening (ranking and selection of dangerous locations). Using the City of Bellevue as an application environment, different safety metrics are automatically generated in the platform such as traffic exposure metrics, frequency of speeding events, and conflict rates. Using Bellevue’s camera network, the proposed approach is demonstrated using a sample of 40 cameras and intersections. The results and platform provide a proactive tool that can constantly look for dangerous locations and risk contributing factors. This paper provides the details of the proposed approach and the results of its implementation. Directions for future work are also discussed.
Autonomous Real-Time Multiple Vehicles Detection and Tracking System
Enas Abu Lebdeh, Yarmouk UniversityShow Abstract
Bashar Awad, Yarmouk University
Ahmad Alomari, Yarmouk University
Mohammed Hawamdeh, Yarmouk University
Mohammad Karasneh, Yarmouk University
Sara Al-Qudah, Yarmouk University
Yara Ajjawi, Yarmouk University
Several traffic studies require vehicle counting in peak hours and during the day with detailed classification and tracking, which exhaust human time and effort, especially at intersections. Manual efforts mainly collect the necessary traffic demand data live in the field or from video records with an extended data manipulation process. Alternative solutions are computer-based systems that efficiently perform human tasks with less time and effort, and these systems vary in their function and performance. This paper proposed a full computer-based system that detects, tracks, and computes related statistics in real-time and maximum utilization of available resources, such as public road surveillance cameras. This work's main contribution is the effectiveness of gathering different computer vision algorithms to achieve high accuracy performance during the real-time streaming of road cameras. The experiments confirm the system performance by achieving, on average, 93.2% as a success rate. The novel addition in this work is that detections, point extractions, matching, tracking, and classification were implemented in a single system that guarantees real-time execution, high accuracy output, and utilizes the available infrastructure. The system overcomes the varying in the light through day and night, and between cloudy and shining weather. Also, it recovers hidden vehicles and the changing in view for each vehicle through its movement. The proposed approach efficiently and partially gathers some mechanisms mentioned above in one system.
Cooperative Ramp Merging In Mixed Traffic: Closed-loop Optimal Control And Real Time Computing
Ziyan Gao, Southwest Jiaotong UniversityShow Abstract
Zheyi Li, Southwest Jiaotong University
Tianyu Huang, Southwest University
Zhanbo Sun (firstname.lastname@example.org), Southwest Jiaotong University
The paper focuses on the problem of decision-making in mixed traffic with conventional human-operated vehicles (HVs) and connected automated vehicles (CAVs). To address the stochasticity in HV behaviors, the paper applies a closed-loop optimal control approach to extend the previously developed deterministic cooperative decision-making for mixed traffic (Deterministic-CDMMT) framework. The Stochastic-CDMMT framework is illustrated using a ramp-merging example. The proposed framework can be described as an optimization problem in which merge sequencing and trajectory design are embedded in a bi-level model predictive control (MPC) framework. The upper-level merge sequencing problem is solved using a dynamic-programming-based solution approach. Aiming at real-time computing, three solution methods, including dynamic programming (DP), sequential quadratic programming (SQP), and bang-bang control are proposed to solve the lower-level multi-object trajectory design problem. Simulation results show that compared to open-loop control, the proposed MPC framework can well tackle the uncertainty in HVs’ behavior. In addition, the solution approach offers a millisecond-level calculation time, which can potentially satisfy the real-time computational needs.
A Markov Decision Process Framework to Incorporate Network-Level Data in Motion Planning for Connected and Automated Vehicles
Xiangguo Liu (email@example.com), Northwestern UniversityShow Abstract
Neda Masoud, University of Michigan
Qi Zhu, Northwestern University
Anahita Khojandi, University of Tennessee, Knoxville
Autonomy and connectivity are expected to enhance safety and improve fuel efficiency in transportation systems. While connected vehicle-enabled technologies, such as coordinated cruise control, have been able to improve vehicle motion planning by incorporating information beyond the line of sight of vehicles, their benefits are limited by the current short-sighted planning strategies that only utilize local information. In this paper, we propose a framework that devises vehicle trajectories by coupling a locally-optimal motion planner with a Markov decision process model that can capture network-level information. By optimizing for a combined short- and long-term fuel and time cost, our proposed framework can guarantee safety and minimize the generalized cost of an entire trip. To showcase the benefits of incorporating network-level data when devising trajectories, we conduct a comprehensive simulation study in two experimental settings, namely a straight highway with on- and off-ramps, and a small network with route choice. The simulation results indicate that further statistically significant efficiency can be obtained for the subject vehicle and its surrounding vehicles in different traffic states under all experimental settings.
Utilizing Traffic Disturbance Metrics to Estimate and Predict Freeway Traffic Breakdown and Safety Events
Leila Azizi, Central Transportation Planning Staff (CTPS)Show Abstract
Mohammed Hadi, Florida International University
The introduction of connected vehicles, connected and automated vehicles, and advanced infrastructure sensors will allow the collection of microscopic metrics that can be used for better estimation and prediction of traffic performance. This study examines the use of disturbance metrics in combination with the usually used macroscopic metrics for the estimation of traffic safety and mobility. The utilized disturbance metrics are the number of oscillations and a measure of disturbance durations in terms of the time exposed time–to–collision (TET). The study investigates utilizing the disturbance metrics in data clustering for better off-line categorization of the traffic states. In addition, the study utilizes machine-learning based classifiers for the recognition and prediction of the traffic state and safety in real-time operations. The study also demonstrated that the investigated disturbance metrics are significantly related to crash. Thus, this study recommends the use of these metrics as part of decision support tools that support the activation of transportation management strategies to reduce the probability of traffic breakdown, ease traffic disturbances, and reduce the probability of crashes.
Smart Traffic Signs to Support Infrastructure-To-Vehicle Communication in the Rural Settings
Enes Karaaslan, Connected Wise LLCShow Abstract
Burak Sen, University of Central Florida
Tolga Ercan, Connected Wise LLC
Haluk Laman, Connected Wise LLC
James Pol, Federal Highway Administration (FHWA)
Connected automated vehicles are undeniably important for advanced safety of our future transportation. However, assisting these vehicles through infrastructure to vehicle (I2V) communication requires substantial investment in wireless infrastructure. Furthermore, access to the power and fiber optic lines is imperative. Therefore, the availability of this technology will be limited to urban settlements in the near-term plans of the transportation authorities. In order to support I2V in rural and underprivileged areas, this project explored an artificial intelligence embedded on-board machine vision system which essentially generates the same I2V messages that are typically sent from the roadside equipment. The system recognizes the message identifiers placed on the traffic signs and activates the associated message while communicating with the vehicle’s on-board equipment, thereby eliminates the need for a wireless infrastructure in these areas. The proposed affordable device solution uses a long-distance stereo vision system to detect in real-time the roadway entities such as traffic lights, traffic signs, vehicles, cyclists, pedestrians and crossing animals. Then, the real-time object locations are accurately estimated by fusing the depth information from the camera and the surveyed road geometry information obtained from the message signs. Some of the example message applications that were proven to be supported by the proposed system include MapData Message for roadway intersections, Traveler Information Message for work zones, Personal Safety Message for vulnerable road users and red-light violation warning messages.
DISCLAIMER: All information shared in the TRB Annual Meeting Online Program is subject to change without notice. Changes, if necessary, will be updated in the Online Program and this page is the final authority on schedule information.