Adaptive Traffic Control Systemsand future of Traffic Signal Control

Ali G. Eghtedari, Ph.D.NW Regional Traffic Operations EngineerMartin Dedinsky NW Region Signal Operations

FHWA Systems Engineering Approach

Model Systems Engineering Documents:

• Concept of Operations

• System Requirements

• Verification

• Validation

SE documents provide a structure within which you can examine your current or near future operation to assess whether or not adaptive control is likely to address your issues and then decide what type of adaptive control will be right for you.

2

Systems Engineering ApproachModel Systems Engineering Documents:

Concept of Operations• Derive the stakeholder needs that should be accommodated by the proposed system• Define the environment in which the system will operate• Provide scenarios that describe how the system is expected to operate in practical

situations• Provide criteria to be used for validation of the completed system

Define the high-level system concept and justify that it is superior to other alternatives

System Requirements• Describes functional requirements of the system, and the conditions it will perform

under

• Defines the necessary requirements to satisfy the operational needs identified in the Concept of Operations;

• Any other requirements necessary for the system to become fully functionalDoes not define how the system is to be built

3

Systems Engineering ApproachModel Systems Engineering Documents

•Verification Plan - Verifies that every step, if done correctly, will fulfill the requirements.

• Describes how the system will be tested to ensure that it meets the requirements

• Details the location of verification testingFAT > BT > SAT

• Vendor is responsible for developing testing method• Details actions and time frame the vendor needs to

resolve any non-compliant functions

•Validation Plan- Validates that every step, if done correctly, will provide a system that meets user needs.• Describes how the performance of the system will be

measured to determine whether the system requirements have been met

• Agency is responsible for validation testing and data collection with vendor assistance

4

ConOPS

Sys. Req.

Ver.

Val.

Traditional Signal Timing Process;

• Traditional signal timing process is time consuming and expensive

• Requires frequent maintenance and updates – i.e. 2-3 years• Final Assessment is often based on anecdotal and

observational judgment due to cost.

5

Adaptive Traffic Control SystemsWhat is ATCS?

An ATCS usually includes algorithms that adjusts:

• Cycle Length• Splits• Offsets• Phase Sequence

In order to:

• Minimize Delays,• Reduce the number of stops,• Decrease the Travel Time.

Any adaptive traffic control system relies upon good detection of the current conditions in real-time in order to allow a quick and effective response to any changes in the current traffic situation.

6

7

Adaptive Systems Objectives• Balance phase utilization—fair distribution of green• Minimize arrivals on red—improve progressed flow• Minimize queue-time density—serve the most cars waiting the

longest• Minimize combination of stops and delay—delay-offset optimization• Pedestrians, Emergency vehicles & RR, Transit vehicles, Light Rail

crossings, Traffic Gating, interface with adjacent systems operations, and other realities

Operations Options:• Work within existing coordination parameters• Override or ignore controller or system coordination• Provide centralized adaptive operation• Provide localized adaptive operation• Optimization suited to grid networks or arterial streets?

8

Architecture Varieties• Built into central system• Works in parallel to central system• Built into local controller• Separate local-cabinet processor• Replaces field master• Built into field masterInfrastructure:• Separate processor may be required• Varying interface to controllers/systems• Varying communications requirements• Detection

– Existing detection– Special detection

Adaptive Traffic Control SystemsBenefits of ATCS

• Improve performance by adjusting to real time traffic demand

• Adapt to unexpected traffic changes• Increases signal timing lifespan• Captures a rich data set

Where is ATCS Most Effective?• Where unpredictable traffic changes

results in delays or stops that cannot be addressed by conventional signal timing.

• Where frequent and unpredictable changes of demand, events, weather situations, etc. creates major unexpected fluctuations in the system.

Adaptive is not solution to all problems of Traffic Management, AND it does not necessarily solve the capacity problem of over saturated corridors.

9

10

Adaptive Benefits & Degree of Saturation

MOE Delay (vehicle-hours)

V/C 0.7 0.8 0.9 1.0 1.1Intersection

ID#FT* SC* %

Ben*FT SC %

BenFT SC %

BenFT SC %

BenFT SC %

Ben

1 32.1 29.0 10% 43.3 36.2 16% 50.0 43.1 14% 68.8 64.5 6% 166.5124.2 25%

2 31.6 29.0 8% 38.4 35.7 7% 47.9 42.5 11% 66.3 67.3 -1% 157.7134.7 15%

3 28.3 25.6 10% 35.4 33.4 6% 44.0 36.6 17% 54.2 54.5 -1% 91.3 116.5 -28%

4 29.2 27.9 5% 36.0 33.9 6% 47.3 41.5 12% 59.5 61.2 -3% 102.9144.4 -40%

Total 121.2 111.4 153.1139.2 189.2163.7 248.8247.4 518.5519.8

Average Benefit 8% 9% 13% 1% 0%

Delay Benefits from SCOOT on the Corridor at Different Congestion Levels*FT = Fixed Time Control, SC = SCOOT Control, % Ben = Percent benefit from SCOOTUniversity of Utah Study (Jhaveri, Perrin, Martin)

11

Adaptive Benefits & Degree of Saturation

MOE Delay (vehicle-hours)

V/C 0.7 0.8 0.9 1.0 1.1Intersection

ID#FT* SC* %

Ben*FT SC %

BenFT SC %

BenFT SC %

BenFT SC %

Ben

1 32.1 29.0 10% 43.3 36.2 16% 50.0 43.1 14% 68.8 64.5 6% 166.5124.2 25%

2 31.6 29.0 8% 38.4 35.7 7% 47.9 42.5 11% 66.3 67.3 -1% 157.7134.7 15%

3 28.3 25.6 10% 35.4 33.4 6% 44.0 36.6 17% 54.2 54.5 -1% 91.3 116.5 -28%

4 29.2 27.9 5% 36.0 33.9 6% 47.3 41.5 12% 59.5 61.2 -3% 102.9144.4 -40%

Total 121.2 111.4 153.1139.2 189.2163.7 248.8247.4 518.5519.8

Average Benefit 8% 9% 13% 1% 0%

Delay Benefits from SCOOT on the Corridor at Different Congestion Levels*FT = Fixed Time Control, SC = SCOOT Control, % Ben = Percent benefit from SCOOTUniversity of Utah Study (Jhaveri, Perrin, Martin)

12

Adaptive Products & Differences

• ACS Lite • BALANCE• InSync• LA ATCS • MOTION

• RHODES • SCATS • SCOOT • Synchro Green • UTOPIA

Adaptive Traffic Control SystemsImplementations:

13

International Locations*Dublin City Council Ireland SCATSNew Zealand Transport Agency

Auckland, NZSCATS

RTA - New South Wales, Sydney Australia SCATS

UOCTConcepcion, Chile SCATS

VicRoadsVictoria, Australia SCATS

City of Blackpool Council UK SCOOTCity of Red Deer Canada SCOOTCity of Southampton UK SCOOTCity of Toronto Canada SCOOTDerby City Council UK SCOOTGreater Manchester Urban Traffic Control UK SCOOTHalifax Regional Municipality Canada SCOOTHampshire County Council UK SCOOTI Mo TS Siemans Ltd. Beijing, China SCOOT

*NCHRP Synthesis 403 - 2010

Local Installations* Total ATCS TypeCity of Menlo Park CA 32 13 SCATSCity of Sunnyvale CA 128 23 SCATSCity of Gresham OR 130 11 SCATSCity of Longview TX 132 16 ACS LTCity of Ann Arbor MI 150 34 SCOOTTown of Cary NC 150 16 OPACCollier County FL 160 16 SCOOTPasco County FL 220 35 SCATSCity of Chula Vista CA 265 11 SCATSPinellas County FL 370 33RHODESCity of Tucson AZ 375 15RHODESWashington State DOT WA 520 10 SCATSCobb County GA 526 74 SCATSOrange County FL 572 70 SCOOTCity of Minneapolis MN 800 56 SCATSDelaware Department of Transportation DE 850 30 SCATSUtah Department of Transportation UT 1100 16 SCATSCALTRANS — District 7 CA 1350 180 LA ATCRoad Commission for Oakland County MI 1500 650 SCATSCity of Toronto Canada 2,100 340 SCOOTLos Angeles Department of Transportation CA 4,300 3,000 LA ATC

14

SCATS

• Sydney Coordinated Adaptive Traffic System (SCATS) has the largest number of worldwide installations – Over 34,000 intersections under SCATS control and largest number of adaptive signals operated in the U.S.

• The intersection controller sends the information collected by detectors to a central server, which automatically adjusts the traffic signal green time to match the traffic flow.

• The central server periodically monitors network wide traffic flow and adjusts all traffic signals in the network.

SCATS FunctionsMajor features:• Stand alone system – Does not require a separate underlying Control

System• Detectors at each intersection detect vehicles approaching and continuously

analyze traffic flow.

Cycle Length• C.L. is calculated to try and maintain Degree of Saturation of 80-90% on lane

with highest DS; Lower and upper limits are user defined.• Algorithm determines critical node in order to calculate the cycle length

Splits• Varied automatically by up to 4% each cycle• Tries to maintain equal DS on competing approaches, minimums are user

defined

15

16

SCOOT

• SCOOT was originally designed to control dense urban networks, such as large towns and cities in UK by the Transport For London, and now is owned by a consortium of TFL, TRL, Siemens and Imtech (now Dynniq).

• There are over 2000 SCOOT systems worldwide working in large congested cities, small towns and around freeway interchanges. There are tens of SCOOT installations in North America, including Toronto, ON.

• SCOOT continually calculates the required coordination pattern for a group of signals in real time and immediately implements the changes.

• Bus priority, traffic gating, incident detection, on-line saturation occupancy measurement, and vehicle emissions estimates are part of the features of SCOOT MC3 systems.

17

ACS Lite

• Examination by FHWA of the barriers to deployment of adaptive control led to the development of ACS-Lite, a cooperative effort with Siemens, McCain, Peek and Econolite.

• Its approach, aims to improve the quality of coordinated control while retaining existing systems with on-street masters, without the installation of large numbers of detectors or need for central server and communication network.

• Its algorithm gradually adjusts the background TOD plans to adapt to gradual changes in traffic conditions, but does not make real-time adjustments as traffic volumes change.

18

RHODES

• RHODES was developed by researchers at the University of Arizona. It currently operates in Pinellas County, FL, and at several others that are being used as test-beds for further development. It continues to be used in research supported by FHWA.

• RHODES, departs from the traditional cycle length, splits and offset approach. It determines when to change state based on current demand at an intersection and predictions of future arrivals at that intersection.

• It was installed on SR 522 as a test base under FHWA’s contract.

19

SynchroGreen

• Synchro is being used by numerous Traffic Engineers around the country as a simple user friendly package for intersection OPS.

• Trafficware and Naztec now offer SynchroGreen which optimizes traffic signals, considering side-street and pedestrian traffic, as well as Transit Priority, in addition to mainline traffic.

• The primary objective of the SynchroGreen algorithm is to minimize total network delay, while providing reasonable mainline progression bandwidth.

• After optimization of splits at intersections, it steps back and analyses the whole network prior to assigning individual timing to intersections.

• Flexible Modes: Balanced, Progression,and Critical Movement.

20

InSync

• Rhythm Engineering released InSync in 2009 and it has been installed and tested in several locations, including Sammamish and Puyallup in WA.

• The underlying philosophy of InSync abandons the concepts of cycle length and phase sequence. It continually evaluates whether a signal should remain in its current state or move to a different state, based on the known demand of traffic at the intersection and predicted arrivals of platoons from other intersections.

• It’s installation philosophy is to retain the existing traffic signal controller and other equipment, and install additional hardware that does the adaptive calculations and commands the controller.

• Uses topology of communicating individual intersections with the cloud via wireless networks.

Purdue University Traffic Signal Performance Measures

21

Purdue MOEs Hi-Resolution Data at 1/10 Second Resolution:

• Phase Termination• Detector Data Diagram• Pedestrian Actuation• Vehicle AOG - Purdue

Coordination Diagram

Most arrivals are during Red!

Purdue University Traffic Signal Performance Measures

• Purdue Coordination Diagram – Plots multiple data points for each signal cycle on a single graph. Graphs only cover a single phase and plot the time of day the cycle began along the x-axis with the time within the cycle plotted on the y-axis. Data points include vehicle arrivals, start of green, start of yellow and start of red splits. Start of green band for instance can be seen as the green line in the above figure.

22

Purdue MOEs Hi-Resolution Data at 1/10 Second Resolution:

• Phase Termination• Detector Data Diagram• Pedestrian Actuation• Vehicle AOG - Purdue

Coordination Diagram

Most arrivals are during Green!

23

Traffic Signal Performance Measures

Split FailuresNumber of split failures during a

defined period that occur on each phase.

Percentage of Phases with Pedestrians

Percentage of signal cycles in a period that have a pedestrian

activation.

24

Intelight & MaxAdapt• MaxAdapt operates in a distributed architecture running on Intelight

ATC controllers and using peer-to-peer communications. • There is no requirement for a central system module to support the

adaptive operations. • The arterial based adaptive algorithm incorporates the high

resolution data collected by the controller and is capable of adjusting cycle length, splits and offsets.

Requirements:• Communications - IP based to support the peer-to-peer

communications.• Detection - MaxAdapt requires mainline detection approximately

400-600 feet from the stop bar at each location. • MaxAdapt will also require stop bar detection for each approach in

order to optimize splits.

25

V2V and V2i Systems!

26

Moving Forward

• Although most adaptive algorithms belong to the past, embrace new technology, consider open source.

– Think out of the box, what does it really mean to say that we want a system that is open source?

• Consider the Client/Server relation - Thin or Thick.

• Consider the scalability and flexibility.

• Focus on Performance Measures as described by Purdue U.

• Don’t get locked into proprietary products.

• Attention to detail - most vendors are good but not most products are.

27

Moving Forward

• Use Adaptive where it is appropriate.

• Select a location where there is a know issue that can be addressed with adaptive.

• Assess, evaluate, and validate the outcome.

• Be aware of staff capabilities and expectations, turn overs, and kind of support we need from the vendor to be able to deliver.

• Effective signal timing is an expensive asset which requires resources to develop and maintain.

• Highly technical subjects such as adaptive signal control are unmaintainable with loss of “institutional knowledge” as we gradually lose capable staff.

Questions?

28

State Traffic Engineers Meeting - May 2016

29