Objective
The objective of this project is to develop traffic signal state transition logic that innovatively employs sensor information. The logic will serve to improve the safety and mobility of vehicles, pedestrians, trains, and light rail transit. The project is not intended to develop a fully implementable product but to prove the concept and justify its further development for implementation.
Status
The revised final report has been delivered.
Product Availability
Several papers have been published in the Transportation Research Record based on this research (all use PDF):
Evaluation of Lane-by-Lane Vehicle Detection for Actuated Controllers Serving Multilane Approaches (2.1 MB)
Evaluation of Flow-Based Traffic Signal Control Using Advanced Detection Concepts (2.6 MB)
Traffic Signal Phase Truncation in Event of Traffic Flow Restriction (0.9 MB)
Modeling Traffic Signal Operations with Precedence Graphs (0.6 MB)
Decision Model for Priority Control of Traffic Signals (0.4 MB)
Advance Preempt with Gate-Down Confirmation: Solution for Preempt Trap (0.6 MB)
Noncoordinated Phases in Coordinated Traffic Signal System: Evaluation of Alternative Permissive Periods on Performance (0.3 MB)
Background
An intersection traffic signal control system can be described as a finite state machine. A finite state machine is a device that stores its state (or status) at a given time and changes states or takes an action when specific, predefined events take place. The transition function of the finite state machine defines the next state of the machine based on the current state and the event. The logic that defines the transition function of an intersection traffic signal controller has changed little over the past 40 years.
Until recently, the only type of traffic information available to the signal controller was whether a vehicle was on top of a detector. Current detection equipment can provide much more information such as queue length, speed, and vehicle classification that could be beneficial in determining when state transitions should be made. Other types of sensors (e.g., pavement surface condition) could also provide useful information. This enhanced information cannot currently be used by the state transition logic in the controller because of the limitations of the interface standard between the controller and the detector.
One area that improved state transition logic could help is mitigating dilemma zones. Drivers are in the dilemma zone if, when they see the yellow indication, they lack adequate distance to stop before the intersection but are too far away to enter the intersection before the red indication. Consideration of vehicle speeds, classification, and location could reduce the number of drivers caught in the dilemma zone.
In addition to the normal operation of the signal, the state transition logic must be able to handle special events such as trains and light rail transit crossing a roadway near the intersection. Preemption of the normal operation must be done quickly and safely in a fail-safe manner to avoid vehicle-train collisions.
Tasks
Phase I
Task 1. Analyze, describe, and critique pertinent domestic and international literature, on the basis of applicability, conclusiveness of findings, and usefulness for the development of traffic signal state transition logic. Review ongoing related research. Key topics are traffic signal state transition logic, detection techniques and technologies, dilemma zone treatments, and highway-railroad grade crossing warning systems.
Task 2. Evaluate current typical signal operations (particularly normal operation, railroad preemption, critical intersection control, and dilemma zone mitigation) to identify the limitations of existing signal state transition logic. Define metrics for assessing how safely and efficiently the intersection operates.
Task 3. Describe the information available from current detection technologies and implementation techniques that could be beneficially used in signal state transition logic. Postulate the information that is likely to be available in the near and more distant future.
Task 4. Brainstorm and describe multiple candidate constructs for signal state transition logic. Assess the candidates' merits and the likelihood that the necessary sensor information will be available around the time that the logic could be implemented in the field. Winnow the candidates to a reasonable number for presentation at the Task 5 workshop.
Task 5. Conduct a workshop of around 20 people to review the results of Tasks 1 through 4. It is expected that the participants will include product developers, researchers, and signal and railroad operators. Solicit and summarize feedback from the participants on the suitability for development of the candidate logic constructs.
Task 6. Refine the candidate logic constructs based on the results of Task 5 and recommend constructs to develop in Task 8. The recommendations should include a description of the challenges and opportunities presented by each construct, both in isolated and coordinated operation and for railroad preemption. The recommendations should include at least one totally new logic construct and some modifications of current logic.
Task 7. Prepare an interim report documenting the findings of Tasks 1 through 6. The report will include an updated, detailed work plan for Tasks 8 and 9. Meet with the panel to review and select the most appropriate signal state transition logic construct(s) for development in Task 8.
Phase II
Task 8. Develop intersection signal state transition logic such that traffic signal vendors can freely adapt it for incorporation into their equipment and researchers can improve it through their research efforts. Describe the sensor systems needed to support the logic.
Task 9. Assess the logic developed in Task 8 to quantify the benefits using the metrics identified in Task 2.
Task 10. Submit a draft final report that documents the entire research effort and includes the documentation of the intersection traffic signal operations logic. A full implementation plan will be required, describing additional work that will be needed and possible incremental implementation paths (see Special Note D).
Task 11. Conduct a workshop on Saturday, January 8, 2005 in conjunction with the TRB Annual Meeting in Washington, D.C. presenting the contents of the draft final report The participants will include representatives from the National Committee on Uniform Traffic Control Devices; the TRB Committees on Traffic Signal Systems, Railroad-Highway Grade Crossings, and Intelligent Transportation Systems; the National Electrical Manufacturers Association; the Institute of Transportation Engineers; the American Association of Railroads, the American Railway Engineering and Maintenance-of-Way Association; the Federal Highway Administration; the Federal Railroad Administration; and the Federal Transit Administration.
Task 12. Revise the draft final report based on the panel's comments. Include a summary of the Task 11 workshop as an appendix.
Phase III
Task 13. Test the logic developed in a variety of situations (i.e., lane by lane operation, railroad grade crossing gate down, phase truncation) and develop a more detailed plan for implementing the results of the research.
Phase IV
Task 14. Develop a scheduling-based concept of coordination. The concept should be simple in application and have good performance when the coordination strategy is changed. The concept will explicitly consider coordination for each direction based on travel time from the upstream approach utilizing advanced (flow based) detection data. The concept will also treat pedestrians as priority requests that do not need to be explicitly accounted for in a background cycle length calculation nor force a separate transition process to be implemented. Develop a working paper for presentation at the Task 15 workshop.
Task 15. Conduct a workshop for a select group of traffic signal control industry representatives and practitioners to present the new concept developed in Task 14. This will be an invitation only event so that participants can actively critique and discuss the issues. Following the workshop, a revised working paper will be submitted to the panel for review.
Task 16. Develop a strategic scheduling algorithm. This task will develop the formulation of the strategic scheduling module. There are two major components of this task. In order to effectively schedule the strategic level decisions, appropriate performance measures must be developed that are reflective of the various priority requests at the strategic and tactical levels. This task will develop potential performance metrics that reflect the need to address coordination, pedestrians, priority vehicles and queue storage. The second component of this task is enhancement of the mathematical model (developed in Tasks 1-13) to include consideration of the various requests. These enhancements include consideration of a priority service interval (for coordination and pedestrian service) and integration with advanced sensor data to consider queue clearance for coordination (progression) and emergency and transit priority requests.
Task 17. Develop a laboratory testing environment. In order to evaluate the new algorithms, some form of software based testing environment is needed. Related research has developed the capability to use Software in the Loop (SIL) using VISSIM (which was used in Phase II of this work) and CORSIM should also be practical. For the early testing of performance measures, it is only necessary to adjust some core logic parameters to test alternative strategies to understand performance issues. Later in the testing process, it will be necessary to run the new coordination logic with multiple intersections and real-time performance measures. The laboratory testing environment developed during this task will be used in the overall assessment described in Task 19.
Task 18. Integrate flow-based sensor data with scheduler. The scheduling algorithm must have appropriate objectives to determine the most appropriate strategy to use for both maintaining coordination and also changing to new coordination strategies. During coordination, the scheduling algorithm needs to evaluate arrivals on green and red indications for each of the coordinated phases to adjust the beginning and ending of the coordinated phase to minimize stops and delay. Likewise, when the strategy needs to change due to higher or lower traffic volumes, the scheduler needs to evaluate whether shortening or extending individual phase times is the appropriate strategy. It may also be appropriate to change phase sequence as part of the changing of strategy.
Task 19. Evaluate the scheduling construct. Evaluation will be done using a software implementation for the priority control logic on CORSIM or VISSIM. Hypothetical networks will allow the team to study the effects of signal spacing and volumes. Real networks typically have a wide variety of signal spacing and interesting demand distributions that also highlight algorithm behaviors.
Task 20. Evaluate and refine the most promising algorithms. It is anticipated that the evaluation will be done at both longer intersection spacing (approximately 1300 feet) where queueing is less of a concern and at closer spacing (approximately 300 feet) where queuing issues are particularly challenging. Volume conditions will vary from lower v/c ratios (0.5) to higher v/c ratios (0.95) to determine performance under both increasing (from 0.7 to 0.95) and decreasing (0.95 to 0.7) traffic volumes. Given that adequate understanding of issues does not exist in oversaturated conditions, it is proposed that such conditions not be included in the evaluation.
Task 21. Submit a final report documenting the entire research effort and revise it based on panel comments.
