ITE Journal – December 2019 - 39

A New Demand/Queue Driven Methodology
In creating a demand/queue driven left turn operation approach,
the base left turn operation for logic processing is utilizing the
flashing yellow arrow. The objective of the methodology in this
research was to determine when to operate protected only or
utilizing the flashing yellow arrow. This intersection was designed
with only advanced or far mainline detection, which utilizes
dilemma zone detection. Left turn detection is typically placed at
the stopline of the intersection, and side street detection has both
stopline and advance detection.
In switching to protected only, each left turn is treated separately.
For the northbound left turn to operate protected only, it looks
at the opposing through detector information, opposing left turn
queue, and northbound left turn queue. For the southbound left
turn to operate protected only, it looks at the opposing through
detector information, opposing left turn queue, and southbound
left turn queue. The left turn queue length can cause obstructions to
driver sight, and is therefore included in the same manner for both
the northbound left turn and southbound left turn, creating safer
scenarios where the left turners can more easily see oncoming traffic.
A queue detection for both left turns is utilized with small
detection zones-approximately 6 feet [ft.] (1.8 meters [m]) in
length, each lane separated. They are set back 200 ft. (61 m) from
the stopbar (queue detector), which amounts to about eight vehicles
(approximately 18 ft. vehicle length and 7 ft. (2.1 m) between
vehicles totaling eight vehicles per lane). The set-back length is
intended to be unique to the geometry of each specific intersection
and should always allow left turning traffic to see opposing through
traffic. There are two delays, totaling 45 seconds (sec.) for the
queue detection. Zero delay would result in rapid switching back
and forth, while a large delay will not switch to protected only
operations fast enough for safety considerations. The 45-sec. delay
was chosen in a virtual controller setting along with video from the
video detection software to ensure the delay was adequate, without
impacting the operations of the traffic signal. As an added measure
of safety, if any of the queue detectors have malfunctioned, the
operation is designed to be protected only until the detector is fixed.
Since going back to the flashing yellow arrow is considered a less
safe operation, the flashing yellow arrow threshold must be met for
120 sec. straight before the flashing yellow arrow can reactivate. For
the flashing yellow arrow to come on in the field, it requires a red
indication from the opposing through phase. As an added measure
of safety, an occupancy value was introduced. If the vehicles are
spaced so closely together, the number of detector actuations are
decreasing in number, but the occupancy value would be increasing
in number.
The opposing through detector was analyzed with closed-circuit
television (CCTV) camera footage at Interstate Highway (IH) 94
and STH 67, historical actuations for both through detectors, and

manual historical intersection count volumes. Figure 2 shows the
correlation between the southbound left turn (SBLT) opposing
through actuations (in 90-sec. intervals) and historical volume at
the intersection. An analysis period close to or equal to the cycle
length is ideal, because it captures the opposing traffic once per
cycle rather than a fraction of the cycle. With field observation,
halfway between the shortest cycle length (80 sec.) and highest cycle
length (100 sec.) provided optimal results.

Figure 2. STH 67 and CTH B (Valley Road) SBLT opposing volume
threshold analysis.
The opposing through actuations increase in conjunction with
the historical volume increase, which would indicate these intersections are below saturation levels. The threshold chosen was 22
actuations and was based on pre-installation data, CCTV footage,
and calibration testing after installation.

Traffic Signal Controller Programming/Coding
The traffic signal controller chosen for this study was the Econolite
ASC/3-1000. The logic processor in the controller programming
made the dynamic flashing yellow arrow coding feasible. All of the
logic statements were made utilizing IF/THEN/ELSE statements,
totaling 49 logic statements. Each left turn operation needed the
ability to go from operating with the flashing yellow arrow to
protected only, and vice versa, when conditions were satisfied and
when the opposing through volume had a red indication. In the
logic statements overlap omits are used, as the flashing yellow
arrows are tied to overlaps, not phases.
The logic statements utilize logic flags that reference different
logic statements and have two alarms programmed in them. Alarm
10 ON refers to the northbound left turn switching to protected
only, and alarm 10 OFF refers to the northbound left turn switching
to flashing yellow arrow operation. Alarm 11 ON/OFF refers to the
southbound left turn in the same manner. The 45-sec. delays for
the queue detectors are in the logic statements, not the detector
programming. The delay in the detector programming is more
utilized for delaying a call into the controller when the corresponding phase is red. The goal of this delay is to not be tied to any
signal indication.
w w w .i t e.or g

D e cember 2019

http://www.ite.org

ITE Journal – December 2019

Table of Contents for the Digital Edition of ITE Journal – December 2019