Project Summary Report 4158-SProject 0-4158: Improving Diamond Interchange Operations Using Advanced

Traffic Controller Features

Authors: Roelof J. Engelbrecht, Steven P. Venglar, P.E., and Zong Z. Tian

Project Summary Report 4158-S – 1 –

Due to financial,environmental, and spatialconstraints, it is becomingincreasingly difficult to improvediamond interchange operationsthrough geometric improvementsalone. An alternative togeometric improvements is tomaximize the use of technology,for example, by better using thecapabilities of currentlydeployed traffic signalcontrollers. Modern traffic signalcontrollers include someadvanced features that are notoften used but have the potentialto improve traffic operations atsignalized diamond interchangesunder certain conditions ifapplied correctly.

What We Did…This research project explored

traffic signal control at diamondinterchanges and in particularhow the features of existingtraffic signal controllers can beused to improve traffic

operations at diamondinterchanges. We developedguidelines to help TexasDepartment of Transportationpersonnel identify and apply theadvanced features available inexisting traffic signal controllersand improve the state of thepractice in diamond interchangecontrol.

We focused on the two trafficsignal controllers that meet thecurrent TxDOT traffic signalcontroller specification:• Eagle Traffic Control Systems’

EPAC300 controller (Figure 1),and

• Naztec Incorporated’s Model980 controller with Version 50software (Figure 2).We investigated the feature sets

of these two controllers andidentified features that could beuseful for diamond interchangecontrol. We then tested thosefeatures to verify that they werecompatible with the diamondcontrol mode that is normally

used for diamond interchangecontrol. Some features were notinvestigated further because theywere incompatible with thediamond control mode.

The remaining features werethen evaluated using atechnique called hardware-in-the-loop traffic simulation,which uses real traffic controlhardware (the actual Eagle andNaztec controllers) to controltraffic in a computerized trafficsimulation running in real time.We used the CORSIMsimulation model andconnected it to a traffic signalcontroller through a TS2controller interface device, asshown in Figure 3. Hardware-in-the-loop traffic simulationallowed us to effectively andaccurately compare controllerfeatures for the followingreasons:• the same traffic conditions

could be reproduced andevaluated under differentcontrol strategies

Summary Report on Improving Diamond InterchangeOperations Using Advanced Controller Features

Figure 1. Eagle EPAC300 Controller Figure 2. Naztec 980 Controller

Project Summary Report 4158-S – 2 –

• extensive measures ofeffectiveness could be used tocompare the performance of thecontrol strategies, and

• the real control hardware added ahigh level of realism.Hardware-in-the-loop simulation

was also used to evaluate theguidelines developed in this projectusing the geometry and demand froma number of real interchangesthroughout Texas.

What We Found…Eight Useful ControllerFeatures

The study identified eightcontroller features that can improvediamond interchange operationsunder certain conditions:• separate intersection mode,• diamond phasing sequence change

by time of day,• conditional service,• dynamic maximum green times,• dynamic split,• volume-density control,• alternate maximum green and

passage times, and• adaptive protected-permissive left

turns.The guidelines developed as part

of this project describe the conceptof operation of each of thesefeatures. They provide detailedapplication guidelines, includingapplicability under differentgeometric and demand conditionsand compatibility with variousdetection technologies. Theguidelines also provide detailedprogramming instructions to guideengineers and technicians inimplementing these features on theEagle and Naztec controllers.

A Useful SeparateIntersection Mode

One of the main findings of thisresearch was a realization of thepotential usefulness of the separateintersection diamond control mode.The separate intersection diamondcontrol mode is not commonly used,

but if applied judiciously it has thepotential to provide more efficientcontrol than the three-phase or four-phase sequences. The separateintersection mode can be used inboth free and coordinated mode.

The free separate intersection modecan significantly reduce stops atinterchanges under low-volumeconditions, especially if the interiorleft turns can operate as permissiveleft turns and steps are taken toreduce the activation of the interiorleft-turn phases.

Figure 4 shows the results of ahardware-in-the-loop simulationperformance evaluation of theBagby/Texas 6 interchange in Waco,Texas. The leftmost bar indicates thepercentage of stops when using thethree-phase sequence, which iscurrently used during low-volumeconditions. The other bars showreduced stops using the free separateintersection mode with different left-turn phasing treatments. Therightmost bar represents the freeseparate intersection mode with theinterior left turns omitted, whichresults in a 34 percent reduction instops compared to the three-phasesequence.

The coordinated separateintersection mode has the potential toprovide more efficient operation thanthe three-phase or four-phasesequence under certain conditions

that can be determined with signaloptimization software such asPASSER III and Synchro. The “ringlag” feature provided by the Eaglecontroller can be used to specify theoffset between the coordinatedphases, allowing the separateintersection mode to provideprogression through the interchangeunder a wide range of geometric anddemand conditions.

A hardware-in-the-loop trafficsimulation of the Briarcrest/Texas 6interchange in Bryan, Texas,indicated that at that interchange thecoordinated separate intersectionmode is more efficient in handlingthe afternoon peak-hour traffic thanthe three-phase mode. Specifically,the analysis indicated that on averagethe separate intersection mode resultsin an overall delay reduction of 2 to9 percent over the three-phase mode.

Video Detection May BeRestrictive

Another important finding of thisresearch is that not all of thecontroller features listed above arecompatible with video detection.Video detection is very suitable forzone detection, but may not besufficiently accurate to serve as inputto controller features that rely on theaccurate point detection of vehicles,such as volume-density control andadaptive protected-permissive left

Figure 3. Hardware-in-the-Loop Simulation Setup

turns. Problems include the irregulargap measurement and occlusion ofvehicles by other, larger vehicles.

Because of the camera angle andresulting parallax error, vehiclelengths observed by the videodetection system are longer thanactual vehicle lengths by a factorthat varies with the height of the rearof the vehicle traversing thedetection zone. Therefore, themeasured gap between vehicles isalways less than the actual gap, andthe difference depends on the heightof the rear of the leading vehicle.Since vehicle heights in the trafficstream vary randomly, measuredgaps will vary randomly, andfunctions that rely on accuratevehicle gap measurement—such asthe adaptive protected-permissiveleft-turn feature that uses measuredgap sizes to determine left-turncapacity—may not operate correctly.

Figure 5 shows the related problemof vehicle occlusion. As shown, it ispossible that a larger vehicle mayobscure a smaller vehicle followingbehind it, resulting in the smallervehicle not actuating the videodetector. This may lead to problemsin controller features that depend onaccurate vehicle counting, such as theadded initial feature of volume-

Project Summary Report 4158-S – 3 –

density control and the adaptiveprotected-permissive left-turn featurethat uses left-turn vehicle counts todetermine left-turn demand.

Increasing distance between thecamera and the detection zoneexacerbates the irregular gapmeasurement and vehicle occlusionproblems if the camera height doesnot increase in the same ratio. Sincethere is typically a maximum heightat which a camera can be mounted,setback detectors often used fordiamond interchange control will bemore susceptible to these types oferrors. One potential solution wouldbe to relocate the camera to a locationcloser to the setback detector.

The ResearchersRecommend…

Since the coordinated separateintersections mode with “ring lag”proved very useful, we recommendthat the TxDOT specification beextended to include the “ring lag” or“individual ring offset” feature that isrequired to implement thecoordinated separate intersectionsmode in the most flexible manner.The Eagle controller alreadyimplements the ring lag feature, butthe Naztec controller does not.

We recommend that the potentialdrawbacks of video detection, asdiscussed above, be taken intoaccount when deciding on thevehicle detection technology to useat new signalized diamondinterchanges, or when a change inexisting detection technology isconsidered.

We recommend that the guidelinesdeveloped in this research be usedwhen reevaluating existing diamondinterchange operation or whendeveloping timing for newinterchanges. The controller featuresidentified in this project should beimplemented, as appropriateaccording to the guidelines, but in allcases, driver expectation and othersafety issues should be taken intoaccount.

Figure 4. Example of Percentage of Stops Under Low-Volume Conditions Based on Bagby/Texas 6 Simulation

Figure 5. Vehicle Occlusion with Video Detection

Camera

Occluded vehicle

0%

10%

20%

30%

40%

50%

60%

70%

Phasing Type

Two-way arterial flow = 230 vphFrontage road flow = 40 to 70 vph

34% Reduction

3-P

has

e M

od

e

Sep

arat

eIn

ters

ecti

on

Mo

de

No

Det

ecto

rS

wit

chin

g

Sep

arat

eIn

ters

ecti

on

Mo

de

Wit

h D

etec

tor

Sw

itch

ing

Sep

arat

eIn

ters

ecti

on

Mo

de

Inte

rio

r L

eft

Tu

rns

Om

itte

d

Per

cen

tag

e o

f S

top

s

Project Summary Report 4158-S – 4 –

DisclaimerThe contents of this report reflect the views of the authors, who are responsible for the facts and accuracy of the data

presented herein. The contents do not necessarily reflect the official views or policies of the Texas Department ofTransportation (TxDOT) or the Federal Highway Administration (FHWA). This report does not constitute a standard,specification, or regulation, nor is it intended for construction, bidding, or permit purposes. Trade or manufacturer namesused in this report do not represent endorsement and appear solely because they are considered essential to the subject ofthe report. This report was prepared by Roelof J. Engelbrecht, Steven P. Venglar (TX-84027), and Zong Z. Tian.

This research is documented in

Related Reports: Report 4158-1, Research Report on Improving Diamond Interchange OperationsUsing Advanced Controller Features

Contact:

Research Supervisor: Roelof J. Engelbrecht, roelof@tamu.edu, (979) 862-3559

Researchers: Steven P. Venglar, P.E., s-venglar@tamu.edu, (512) 467-0946

Zong Z. Tian, z-tian@tamu.edu, (979) 862-2458

Project Director: Kirk Barnes, P.E., TxDOT Bryan Districtkbarnes@dot.state.tx.us, (979) 778-9756

To obtain copies of the reports, contact Dolores Hott, Texas Transportation Institute, Information &Technology Exchange Center, (979) 845-4853, or e-mail d-hott@tamu.edu. See our on-line catalog athttp://tti.tamu.edu.

For More Details . . .

The product resulting from this research was a guideline for use of advanced traffic control features, includingmatching of features with conditions, programming changes, and hardware needs. Advanced features of EagleEPAC300 and Naztec 980 controllers were evaluated using CORSIM simulation and field testing. The guidelineswere published with report 4158-1. The guidelines have been incorporated into the “Diamond Interchange SignalTiming Workshop” that was developed as part of Project 5-1439-01: Transportation Research ImplementationConsortium for Operations and Management. The workshop will be presented at a number of locations in Texasover the next 3 years.

For more information, please contact Wade Odell, P.E., RTI Research Engineer at (512) 302-2363 orwodell@dot.state.tx.us.

YOUR INVOLVEMENT IS WELCOME . . .

TxDOT Implementation StatusApril 2002

ITEC.PSR0201.0802.580