Chapter 8

Intersection Control
an d Design

1. INTRODUCTION

C h ap ie r 7 d e a lt w ith th e basic concepts relatin g to th e capacity and level of service of

tran sp o rta tio n facilities w here th ere was u n in te rru p te d traffic flow (e.g., freew ays and
rural tw o-lane highw ays). O n e of th e m ost logical ways o f continuing with that ch ap ter
would have b een !•- discuss the capacity of intersections. H ow ever, it wr.s felt th at before
such an a tte m p t was m ade, it w ould b e b est to u n d ersta n d the fun d am en tals of how in ­
tersectio n s are co n tro lled and d r ' ;ned. ..M this is f . w hat w e describe in this
chapter. The capacity and level o f service of intersections are taken up in C h a p ter 9.

2. TYPES OF IN TERSECTIO N S

Intersectio n s are an inevitable p a rt of any stre e t system . D riving aro u n d any city, one
notices that a large m ajority o f u rb a n streets sh are an intersection, w 'ic re drivers can
dccide w h eth er to jio straight o r tu rn on to a n o th e r street. A road o r street intersection
can be defined as th e general a re a w here tw o o r m oio ro ad s join o r cross, irc li'jm g th e
roadw ay and roadside facilities for traffic m o v em en t w ithin if (A A S H T O , 1994).
B ecause an in tersectio n has to be sh ared by everybody w an t:ng to use **. it needs
to be designed with great care, taM g g in to consideration efficiency, safe ty , speeo, cost of
o p eratio n , and cap acity. T h e actffiP fraftic m o v em ent an d its seqv.er.t r carfTieTiandled
By various m eans, d ep en d in g on th e type of in te^f-civ m n eed ed (A A S H T O , 1994).
in g en eral, 'be-ve a rc th re e ty p es o f in tersectio n s: (1) in te rsec tio n a t grad e,
(2) grade sep aratio n s w ithout ram ps, and (3) in terchanges. T h e com m on intersection at
grade is o n e w h ere tw o or m o re highways jo in , with each highw ay radiating lro m an
intersectio n and fo rm in g p a rt o f it. T h ese a p p ro ach es are re fe rre d to -as intersection
legs. Such intersections have th eir ow n lim itations and use. E xam ples of at-grade in te r­
sections are ^ " w r in Figure 8-1. W hen it is necessary to accom m odate high volum es of

271
272 Intersection Control and Design Chap. 8 J

y V.
'vl

UNCHANNELIZED T FLARED T T WITH TURNING ROADWAYS

Y WITH TURNING ROADWAYS

3 - LEG IN T E R S E C T IO N S

J-H'1
-H
T

FLARED CHANNELIZED

4 -L E G IN T E R S E C T IO N S
■

M U L T IL E G IN T E R S E C T IO N

Figure 8-1 Examples of At-Grade Intersections.

traffic safely and efficiently through intersections, one resorts to through traffic lanes
separated in grade, and this is generally referred to as interchange The basic types
of interchanges are shown in Figure 8-2. W hen two highways or streets cross each
other at a different grade, with no connections, the arrangem ent is referred to as a
grade separation.
D etails of the geom etric design of at-grade intersections and interchanges are
given in A Policy on Geometric Design o f Highways and Streets published by the A m er­
ican Association of State Highway and Transportation Officials (1990,1994), and this
book should be consulted for a deeper insight into this im portant facet of design.
Sec. 6 Types of Intersection Controls 2?D

W ITH O tV ISIO N & L IS L A M ) AND TU RM UG ROACWAYS

(a)

(b)

Figure 8*4 T y p ical E x a m p le s o f C h a n n e liz e d in te rs e c tio n (A A S H T O , 1994).

 . , 274 Intersection Control and Design Chap. 8

1. H um an factors* such as driving habits and decision and reaction times

2. Traffic considerations, such as capacities and turning'm ovem ents, vehicle speeds,
and size and distribution of vehicles
3. Physical elements, such as characteristics and use of abutting property, sight dis­
tance, and geometric features
4. Economics factors, such as costs and benefits and energy consumption
.

Insofar as interchanges are concerned, their type and design are influenced by
many factors, such as highway classification, character and composition of traffic, design
speed, and degree of access control. Interchanges are high-cost facilities, and because of
the wide variety of site conditions, traffic volumes, and interchange layouts, the war­
rants that justify an interchange may differ at each location. A A SH TO (1990, Chap. IX)
provides details regarding grade separations and interchanges. The bottom line when
considering adoption of an interchange is: Can the cost of an interchange be justified? ;

4. TRAFFIC CONTROL DEVICES

Traffic control devices include signs, movable barriers, and signals. A ll these can be
used alone or in combination if necessary. They arc the piim ary m eans of regulating,
vvarning, or guiding traffic, on all streets and highways. Traffic control devices strive to
provide safe and efficient functioning of intersections by separating conflicting vehicle
stream s in tim e. In other words, the right-of-way through an intersection, during a
' given period, is assigned to one or several streams of traffic. For example, yield or stop
signs assign priority to particular traffic streams relative to other streams at the same
intersection. A four-way stop sign establishes a rough first-come, first-served traffic
control for intersections with light traffic.
The M anual on Uniform Traffic Control Devices (M UTCD) (FH W A , 1988) sets
forth the principles that govern the design and usage of traffic control devices for all
streets and highways open to public travel, regardless of type of class or the govern-
. m ental agency having jurisdiction.
Traffic signs and markings are used for establishing a street or highway system
. that is clearly up'W ctood by ’ts use"4’: drivers and pedestrians. Specifically, traffic signs
anu markings fulfill the following purposes: the regulation of traffic (e.g., speed limits),
turn prohibition, alerting and warning drivers and pedestrians regarding roadway con-
: ditions, and guiding traffic along appropriate routes to reach trip destinations through
signs and markings. These purposes apply to all control devices, including signals,
markings, and channelizations. N aturally, to be effective, control devices m ust m eet
the following basic requirem ents:

1. Fulfill a need
.
2 Comm and attention
3. Convey a clear, simple meaning
 Sec. 5 Conflict Areas at Intersections 275

4. Command respect of road users

5. Give adequate time for proper response

The following criteria should be applied to ensure that these requirem ents are
met. Signs should be of a proper design and should be placed properly and appropri­
ately. They should be operated consistently and m aintained routinely. Finally, there
m ust be uniform ity in application, so th at recognition and understanding of these
devices is easy and unambiguous.
Traffic signs fall into four broad areas of functional classification according to use:
m

1. Regulatory signs are used to impose legal restrictions applicable to particular

locations. They inform drivers of certain laws and regulations, the violation of
which constitutes a misdemeanor. There are four principal groups of regulatory
signs, excluding those lor pedestrians. They are right-of-way signs, the most com­
mon being stop and yield signs; speed signs; m ovem ent signs, such as turning or
one-way signs; and parking signs.
2. Warning signs are used to call attention to hazardous conditions, actual or poten­
tial, that would otherwise not be readily apparent. Such signs require caution on
the part of the driver and may call for a speed reduction or other m aneuver.
Typical conditions where warning signs are used include highway construction
zones a^d appr^achc^ +o intersections, merging areas, pedestrian crossings, and
school zones.
3. Guide or informational signs provide directions to drivers and 1o various destina­
tions. These are placed far enough ahead of intersections and interchanges to
allow adequate time for drivers to make their routing decisions.
4. Directional signs on high-speed high /ay1’ a .e used at interchanges associated
with freeways.

CONFLICT AREAS AT INTERSECTIONS

Figure 8-3 shows vehicle streams and the merging, diverging, and crossing maneuvers
for a simple four-leg intersection, and for a m ore complicated staggered intersection.
Such diagrams are useful because the num ber and type of conflicts may indicate the
accident potential of an intersection. In the case of a regular two-lane, two-way, four-
leg intersection there are 16 potential crossing conflict points, eight merging and eight
diverging conflict points. The staggered T-intersection shown in the figure serves about
the same function as the four-leg intersection, and consists of only six potential crossing
conflict points, three 'diverging and three merging conflict points. Is the staggered
T-intersection, therefore, superior to the four-leg intersection? N ot really. T here are
several other factors that play an im portant part in deciding the merits of adopting a
particular type or design of intersection for a specific site.
 276 Intersection Control and Design Chap. 8

# 1 6 crossing con flicts -

O 8 merging con flicts
© 8 diverging con flicts

• 6 crossing conflic ts
O 6 merging conflicts
Q 6 diverging conflicts

0 -at)".,, Q

i'igur® 8-3 Vehicle Streams and the

Merging, Diverging, and Crossing
Vehicle streams at a three-way intersection Maneuvers (baiter, 19/4).

TYPES OF INTERSECTION CONTROLS

There are at ieiis! six principal ways of controlling traffic at intersections, depending on
the type of intersection and the volume of traffic in each of the vehicie streams. The
MUTC'D (FHW A, 1988) provides guidelines for adopting any particular type of inter­
section control, in the form of warrants.

6.1 Stop Signs

Stop signs are w arranted at intersections under the following conditions:

1. Intersection of a less im portant road with a main road, where application of the
norm al r i g h t - o f ^ y rule is aAflfrbly hazardous
2. Intersection of a county road, city street, or township road with a state highway
3. Street entering a through highway or street
Sec. 6 Types of intersection Controls 277

4. Unsignalized intersection in a signalized area

5. Unsignalizcd intersection where a combination of high speed, restricted view, and
serious accident record indicates a need for control by the stop sign

M ultiway (fourway or all-way) stops can be used as a safety m easure at some

locations where the volume on the intersecting roads is approximately equal and the
following conditions exist:

1. A n accident problem , as indicated by five or_™ore reported accidents in a

12-month period, which may be corrected by a multiway stop installation.
2. (a) The total vehicular volilme entering the intersection from all approaches aver­
ages at least 500 vehicles per hour for any 8 hours of an average day, and (b) the
com bined vehicular and pedestrian volum e from the m inor street or highway
averages at least 200 units per hour for the same 8 hours, with an average delay to
minor street vehicular traffic of at least 30 seconds per vehicle during the maxi­
mum hour, but (c) when the 85th-percentile approach speed of the m ajor street
traffic exceeds 40 mph, the m inim um vehicular volum e w arrant is 70% of the
foregoing requirements.
3. W here traffic signals are w arranted, the multiway stop control can be used as an
interim m easure while arrangem ents are being made for installation of the signal.

6.2 Yield Signs

Yield signs are established as follows:

1. On a m inor road at the entrance to a n intersection wk«en it is necessary to assign

the right-of-way to the major road, but w here a stop is not necessary at all times,
and where the safe approach speed on the m inor road exceeds 10 mph
2. O n the entrance ram p to an expressway, where an adequate acceleration lane is
not provided
3. W here there is a separate or channelized right-turn lane w ithout an adequate
acceleration lane
4. A t any intersection where a problem can be possibly corrected by a yield sign
installation
5. W ithin an intersection with a divided highway, where a stop sign is present at the
entrance to the first roadway, and further control is necessary at the entrance to
the second roadway. M edian width between roadways must exceed 30 ft.

6.3 Intersection Channelization

Channelization is the separation or regulation of conflicting traffic movements into def­

inite paths of travel by traffic islands or f aven.rnt m a i k i^ , .o facilitate the safe and
orderly movements of both vehicles and pedestrians. Proper channelization increases
278 Intersection Control and Design Chap. 8

capacity, improves safety, provides maximum convenience, ana instills driver confi­
dence. Channelization is frequently used along with stop or yield signs or at signalized
intersections.
Some basic principles to help design channelized intersections are as follows:

1. M otorists should be provided with channel lines that are easy to follow.
2. Sudden and sharp reverse curves should be avoided.
3. A reas of vehicle conflict should be reduced as mucu ns possible.
m
4. Traffic stream s that cross w ithout merging and weaving should intersect at or
neat' right angles.
5. Islands should be carefully selected and be as few as possible.
6. O verchannelization should be avoided, as it has proved to be counterproductive.

Figure 8-4 gives typical examples of channelized intersections. The Intersection

Channelization Design Guide (TRB, 1985) and A A SH T O 0 9 9 4 ) provide further
details on channelization.

6.4 Rotaries a n d Roundabouts

Rotaries and roundabouts are channelized intersections comprising a central circle sur­
rounded by a one-way roadway. The basic difference betw een rotaries and round­
abouts is that rotaries are generally signalized (as in W ashington, D.C.), w hereas
roundabouts are not. Naturally, in the case of ro u n d a b ^ ls , eiilering Uaffic yields to
traffic already within.
R oundabouts generally have good safety records and traffic does not have to stop
when traffic volumes ! >w. A vw V de^gned roundabout should deflect the path of
vehicles passing through an intersection by the use of a sufficiently large central island,
properly designed approach islands, and staggering the alignment of entries and exits
(Figure 8-1).

6.5 Uncontrolled Intersections

W here an intersection has no control device w hatsoever, the operator of a vehicle

approaching an intersection must be able to perceive a hazard in sufficient time to alter
the vehicle’s speed, as necessary., before reaching the intersection. The time needed to
start decelerating is the driver’s perception and reaction time and may be assumed to
be 2.Q seconds. In ad d itid lC fl^ fn v eF rieed s to begJrTbraFmg some distance from the
intersection. This distance from the intersection, where » driver can first see a vehicle
approaching on the intersecting road, is that which is traversed during 2.0 seconds for
perception and reaction, plus an additional 1.0 second to actuate braking or to acceler­
ate to regulate speed. By referring to Figure[8:5[ the sight triangle is determ ined by the
minin'ium distances along the road. For instance, if highway A has a speed limit of
50'm ph and highway ‘f w ..one of ?-j m\ I., it would require an unobstructed sight tri­
angle, with legs extending at least 220 ft and .130 ft, respectively, from the intersection,
280 Intersection Control and Design Chap. 8

SIGHT TRIANGLE

a case i an
NO CONTROL OR YIELD CONTROL ON MINOR ROAD

¥
O) IS)

8CASE m
STOP CONTROL ON MINOR ROAD

Figure 8-5 Sight Triangle (AASHTO, 1994).

based merely on the average distance traveled in 3 seconds. These minimum distances
will perm it a vehicle on either road to change speeds before reaching the intersection,
but this fact by itself does not imply that the intersection is safe.
; There can be potential danger to vehicle operators on such intersections, espe­
cially when a succession of vehicles are approaching the intersection, when time is suf­
ficient to avoid only a single vehicle. Because the distance covered in 3 seconds ranges
from 70% of the safe ai^p^ing disu..ace at 20 mph to only 36% at 70 mph, the use of
sight triangles for design purposes must be approached with caution.
 Sec. 7 Traffic Signals 281

A safer design for such intersections should allow drivers on both highways to see
the intersection and traffic in sufficient time to stop the vehicle before reaching the
intersection. The safe stopping distances in this case are the same as those used for
designing any other section of highway.

6.6 Traffic Signal Devices

O ne of the most im portant and effective methods of controlling traffic at intersec­

tion is the use of traffic signals. The traffic sig.iai is an eiecincally timed device that
assigns the right-of-way to one or m ore traffic stream s so that these traffic streams can
pass through the intersection safely and efficiently. Traffic signals are appropriate for
minimizing:

1. Excessive delays at slop signs and yield signs

2. Problems caused by turning movements
3. Angle and side collisions
4. Pedestrian accidents

B ecause traffic signaling devices are so im portant, the rest of this chapter is
devoted to describing their design and application.

TRAFFIC SIGNALS

All pow er-operated devices (except signs) for regulating, directing, or warning
m otorists or pedestrians are classified as t i «,Tfic signals. Tiub ^ a i o n begins with a basic
set of definitions connected with traffic signals and intersections and goes on to detail
several m ethods of designing signal timing. The various sections provided on this sub­
ject are necessarily just th~ basics. Further details can be found in several references
provided at the end of this chapter.

7.1 The Purposes of Traffic Signals

In general, a traffic signal is installed at an intersection:

• To improve overall safety

• To decrease average travel tim e through an intersection, and consequently
increase capacity
• To equalize the quality of service for all or most traffic streams

A lthough traffic signals are installed on the basis of w arrants, justification for
their installation must be m ade in terms of safety, travel times, equity, pollution, and so
H; on. M ore will be said about w arrants in a I: ter a c tio n .
Amont* the m»in advantages of traffic signals over sign control are the positive
1" guidance that they provide to vehiclc operators and pedestrians, which leaves less room
282 Intersection Control and Design Chap. 8

for erroneous judgm ents on the part of drivers; flexibility, in the sense that allocation of
right-of-way can be responsive to change in traffic flow; ability to assign priority treat­
m ent to some movements or vehicles; feasibility of coordinated control along streets or
in' aiea networks; and provision for continuous flow of a platoon of traffic through
proper coordination at a specified speed along a given route. On the other hand, it has
been observed that poorly designed traffic signals can cause increased accident fre­
quency, excessive delay for vehicles on certain approaches, forcing motorists to adopt
circuitous routes, and driver irritation.
#

7.2 Definitions Pertaining to Intersections and Traffic Signals

A num ber of terms used in this and other chapters need to be defined. They have been
extracted from the Traffic Engineering H andbooL (Plinv, 1992), and the H ighway
Capacity Manual (TRB, 1994).

1. Cycle (cycle length or cycle time): any com plete sequences of signal indications.
^ 2. Phase (signal phase): the p art of a cycle allocated to any com bination of traffic
movements receiving right-of-way simultaneously during one or m ore intervals.
3. Interval: the part or parts of the signal cycle during which signal indications do
not change. ~ ‘
4. Offset: the time lapse, in seconds, betw een the beginning of a green phase at the
intersection and the beginning of a green phase at the next intersection.
5. Intergreen (clearance interval): the time betw een the end of a green indication for
one phase and the beginning of a green indication for another (Figure 8-6).

Green Yeilow Red

: Street A

Hud Green

Street R

‘ntergreen \~*~

Green Yellow Red

y?777)
Street A
S' / / / / w m
--- ^ ------- All Red \ _
Red (-
C inetfW Street 8

-j Intergreen

-Timp -
Figure 8-6 Intergreen.
 Sec. 7 Traffic Signals 283

6. All-red. interval: the display lime of a red indication for all approaches. In some
cases, an all-red interval is used exclusively for pedestrians crossing very wide
intersections (Figure 8 6 ). ~
7. Peak-hour factor (PHF'): in the case of street intersections, the ratio of the num ­
ber of vehicles entering the intersection during the peak hour to four times the
num ber of vehicles entering during the peak 15-minute period. In the absence of
field information a PH F value of 0 ^ ) m a y be assumed, in which case, the traffic
flow during the 15-minute period - N / ( 4 X 0.85) = Q.294N, or approximately
0.3N, where N is the peak-hour flow.
- 8. Average departure headways: observations made by G reenshields et al. (1947)
show that for green intervals of 20 t&3Q Seconds, visage headway per vehi­
cle is about 2.5 seconds.
^ 9. Passenger-car equivalents (PCEs): to account for the adverse effects of commer-
a \ 'f ^ cial vehicles and turning m ovem ents on startup time (or average head way), it is
* \ ^ o-'j' ' customary to convert actual flows (given in mixed vehicles per hour) to an equiv-
v « alent volume in straight-through passenger cars. Buses and trucks arc assumed to
J1, y ^ T b e 1.5 PCE, and left-turn vehicles are approxima1iIy~equal to 1.6 PCE.

5 o 0 ' f J
,> ' In addition, the following terms will be used from time to time in this and subse-
. /v quent chapters (TRB, 1994):
/
Approach: the portion of an intersection leg that is used by traffic approaching
the intersection * '
Cajxi/ity: the maximum num ber of vehicles that has a reasonable expectation of
passing over a given roadw ay or section of roadw?ay in one direction during a
giVen time period under prevailing roadway and traffic conditions
- Critical volume: a volume (or combination of volumes) for a given street th at pro-
duces the greatest utilization of capacity (e.g.. needs the greatesT~green tim e) for
that street, given in terms of passenger cars or mixed vehicles per hour per lane
Delay: the stopped tim e delay per approach vehicle (in seconds per vehicle)
Green time: the length of green phase plus its change interval, in seconds
^ Green ratio: the ratio of effective green time to the cycle length
H ourly volume: the num ber of mixed vehicles that pass a given section of a lane
or roadway during a time period of an hour
Level o f service: a m easure of the mobility characteristics of an intersection, as
determ ined by vehicle delay, and a secondary factor, volume/capacity ratio
Local bus: a bus having a scheduled stop at an intersection
Passenger-car volume: volume expressed in terms of passenger cars, following the
application of passenger-car equivalency factor to vehicular volumes
K Period volume: a design volume, based on the flow rate within the peak 15 min-
jj& utes of the hour, converted to an equivalent h o u rlv volume
284 Intersection Control and Design Chap. 8

Through bus: a bus not having a designated stop at the intersection under analysis
Truck: a vehicle having six or more wheels (tires) on the pavem ent

7.3 Components of a Signal System

A signal installation consists of illum inated displays and a controlling mechanism.^It

may also include " ;^us veh iJo -d e’. cting devices or some other means for activation
by dem and (such as a push button for pedestrians desiring to cross a street).
Displays, or indications, are grouped into signal faces, each of which controls one
or m ore traffic streams arriving from the same direction. A signal head contains one or
m ore signal faces; it can be m ounted on a post or suspended from a wire.
Signal indications differ by color, shape, and continuity. The colors used are:
green, to give the right-of-way to one or a com bination of traffic streams; red, to pro­
hibit movem ent or to require a stop; amber, to regulate the switching of the right-of-
way from one set of traffic stream s to another or to advise caution. W hen there are
special signals for pedestrians, these are in the forw of illuminated letter messages or
logos. Signal indications can be steady or flashing. As noted before, a flashing red indi­
cation has the same meaning as a stop sign, whereas a flashing am ber allows one to
proceed with caution. The flashing “walk” cautions a pedestrian that a vehicular stream
is concurrently perm itted to cross his or her line of movem ent. The flashing “don’t
w alk” is the equivalent of an am ber indication.
Signal controllers are electrom echanical or electronic devices that regulate the
length and sequence of signal indications at^an intersection. Pretim ed controllers oper­
ate with a fixe'4 - ^ o u n t of t’me 1lotted to specinc traffic m ovem ents in a fixed
sequence; the timing is based on historical flow patterns at an intersection. Traffic-
adjusted controllers are equipped to receive inform ation on traffic flow patterns from
various measuring devices at preset time intervals. This inform ation is used to select
one of several timing schemes stored in the controller’s memory.
Traffic-actuated controllers also use some sensing devices to alter the length
and/or the sequence of signal indications. Unlike traffic-adjusted controllers, however,
they react to arrivals of individual vehicles rather than to changes in aggregate patterns
of traffic at an intersection. Traffic-actuated timing schemes are usually constrained by
specified minimum lengths of green indications for various traffic streams that can be
marginally extended by vehicle arrivals up to specified maxima. Intersections can be
controlled individually; alternatively, a sequence of intersections along a road could be
connected and controlled as a group.
D etectors can be activated by the passage or the presence of a vehicle. A variety
of physical principles are used for detection: pressure, distortion in a magnetic field,
interruption of a light beam, a change in radar wave frequency, a change in inductance
of a conducting loop, video detection using image processing techniques, and so on.
They can be p o rtio n e d abov ~ the r^ad or on, inside, or under the road surface, or sim­
ply on the video screen for image processing. D etectors differ in investm ent and m ain­
tenance cost as well as in reliability.
Sec. 7 Traffic Signals 285

Figure 8-7 Signal Phases.

7.4 Elements of a Signal Timing System

A t signalized intersections, some traffic stream s are allowed to have a simultaneous

right-of-way, while other streams are stopped. A signal phase is a period during which
one or m ore movements concurrently are shown a green indication. Figure 8-7 shows a
phasing diagram for ar> intersection of two streets.
Safety considerations dictate th at a phase may be shared only by those traffic
streams whose paths do not intersect. In practice, however, some conflicts are toler­
ated. In Figure 8-7, pedestrians and right- and left-turning vehicles enjoy concurrent
green. Left turns are often allowed to clear an intersection through gaps in the traffic
streams moving along the same street but in the opposite direction.
A t some intersections, where both pedestrians and turning vehicles are num er­
ous, it may be advantageous to provide a special phase w here all pedestrian m ove­
ments arc perm itted while all vehicles are stopped. Subsequent phases are m eant for
vehicles only.
The timgjbetween the end of a green indication for one phase and the beginning
of a green for another is called intergreen, or a clearance interval. A n am ber indication
is shown through the intergreen period followed by red. This is shown in Figure 8-6.
W hen the computed clearance interval is long, a combination of am ber and an all-red
interval may be used instead. This is also illustrated in Figure 8-6.
Design of signal phases specifies a sequence of various phases following each
other. A signal cycle is a part of that sequence. N ote that safety (conflict avoidance)
and the quality of scrvice are the most im portant factors in designing signals.
A phasing diagram is developed by jointly considering intersection geometry (pri­
marily, the num ber of lanes at each approach) and desired lines of movem ent through
an intersection. Safety is the sole criterion in computing intergreen times, as shown in
what follows. Factors affecting the length of intergreen include safe stopping distances,
approach speeds of vehicles, walking speeds of pedestrians, and pavem ent widths.
 286 Intersection Control and Design Chap. 8

W hen the sum of intergreens over all phases is subtracted from cycle time, what
remains is the total green time per cycle. Selection of green times depends on whether
to minimize the overall average travel time through an intersection, or to equalize
dem and and capacity over a given time period, or to minimize the maximum individual
travel time through an intersection, and so on. Each objective may result in a different
set of cycle times and green indications.

Calculation of the Intergreen Period. A n examination of Figure 8-6 indicates

that the intergreen p r i o r i , -consisting of either an am ber or an am ber plus an all-red
period, is necessary to _alert motorists regarding the change from a green light to a red
light. W hen the am ber indication appears, drivers who are at a distance (from the stop
line)-greater than their stopping distance will be able to stop com fortably. Referring to
Figure 8-8, those drivers who are aearer the stop line thaw thei* safe stop distance will
‘ acceleafcte and clear the intersection. Those who are at or near a safe stop distance
away (the so-called “dilemma zone”) should be able to (1) either stop or (2) accelerate
(when near the stop line) and clear. The path of the vehicle covers a distance
rSL S + W + l (i)
where
S safe stopping distance
W — distance from the stop line until vehicle (rear) is clear
I = length of the vehicle
Therefore, the intergreen time is

S+W +l v W+l
U _|----------------_|-----------------
v r 2f x 8 v
(2)

where
I i?- 3 3 o . C 'L
3U .

v = speed of the vehicle (ft/sec)

g = 32.2 ft/sec2 bj>J'

- -j i U
cu
o4) ■ ^ 2.
CO f>
C _ > j5 \ JlJP ij-*
, ;vvt s \
Street A
I* * ' Xf
cz>
/ i
h- -w- h-
Figure 8-8 Clearance Distances.
Sec. 7 Traffic Signals 287

tr — perception-reaction time (1 to 2.5 sec)

/ = 0.33 (wet pavem ent) to 0.62 (dry pavem ent) at 30 to 40 mph
W and I are m easured in feet. If there is no all-red (a —0),
I =A (3)
that is, intergreen is equal to an am ber indication. Now consider the am ber indication
time based on pedestrian movement, in the same direction as the car, crossing street B.
If there are no pedestrian signals, assume that the last pedestrian will start crossing
exactly as the am ber indication comes on. The pedestrian clearance time is

(4)
where Wj is the street width or up to the median, and vped is the walking speed of pedes­
trians (4 ft/sec). If R i is longer than /,, use R t as .

Example 1
Determ ine the optimal'duration for the yellow phase to eliminate the dilemma zone at an
intersection. It is assumed that the dilemma zone can be eliminated by adjusting only the
yellow-light duration.
Solution
To stop a vehicle before the stop line, the driver has to start braking at a distance of

from the stop line, called the stopjtone. The term tr here is the driver’s reaction time, v is
the travel speed, and a is the deceleration rate.
On the other hand, an adequate yellow phase Y should be provided to allow a vehi­
cle to clear the intersection if the v ’ucle too cloSv, ’ 3 intersection when the yellow
phase starts. The distance to the stop line from the vehicle, which is to clear the intersec­
tion during the yellow phase, is called the go zone, given by
G = v Y - (W + I)
where W is the street width, and I is the vehicle length.
The dilemma zone is the distance between G and S, that is,
D = G - S = v Y - (W + /) - (trv + v l/2a)
To r m im ’^ O we set D = 0, to get

which is the optimal duration for the yellow phase.

Discussion
The purpose of eliminating the dilemma zone is to help drivers set up a clear bench­
mark, or imaginary line, behind which the vehicle should brake, or beyond which the vehi­
cle should go through the intersection, when the yellow phase starts. However, due to
283 Intersection Co.u.ol and Design Chap. 8

different geometric features ot intersections, the dilemma zone of some intersections may
not be totally removed by adjusting only the yellow phase. When there is no all-red
(a = 0), the intergreen time is the same as the amber (yellow) phase.

The Shortest Cycle Based on Pedestrian Requirements. In urbanized areas

of a city where pedestrian circulation is im portant and necessary, it is the general prac­
tice to provide pedestrian control signals. W hen “w alk/don’t w alk” indications are
used, pedestrian phase length might govern the green indication for th at approach.
Pedestrian’? need a totu, n u ^ t o cro ^ an ii^ rs e c tio n of
Z 4- R , seconds (6)
where Z is the initial period during which the “walk” signal is displayed, usually « 7
seconds; and R i is the clearance time for the last pedestrian who starts crossing the
intersection when the pedestrian signal just begins to flash “don’t walk.”
A s part or all of R [ can coincide with the intergreen time, the pedestrian phase
requires only
P ^Z +R -f (7)
where P{ is the pedestrian phase (sec) during the green indication. (Note: It is assumed
that Ri is longer than if 1 is the larger, Pt = Z.) It follows that
Pi = min G t (8)

7,5 Signal Timing for Pretimed Isolated Signals

>
O ver the years, traffic engineers ha'-e used several m ethods for designing pretim ed
isolated signals. In m ore recent times, a num ber of com puter program s have been
developed for rapid design. T hree m anual m ethods of design are presented here:
(1) H om burger and Kell’s method, (2) Pignataro’s method, and (3) W ebster’s method.

Homburger and Kell's Method (Homburger et al., 1988). This m ethod uti­
lizes traffic volumes as the basis for allocating time to approaches, keeping off-peak
cycles as short as possible (40 to 60 seconds). Peak-hour cycles can be longer, favoring
movem ent on the major street. The general procedure is as follows:

1. Select yellow change intervals betw een 3 to 5 seconds for speeds less than 35 mph
to speeds greater than 50 mph.
.
2 D eterm ine the need for additional clearance time u&ing Lq. 5 and also ensure if
an all-red phase is necessary.
v W+ /
Y — tr + - 1----------
2a v
assum ing t h a t 1 J e c e le ra r m r; , j o f 10 ft/sec2.
3. D eterm ine pedestrian clearance times, assuming pedestrian walking speed as
4 ft/sec.
 Sec. 7 Traffic Signals 289

4. C om pute minimum green times. W ith pedestrian signals, the “w alk” period
should be at least 7 seconds.
5. Compute green times based on an approach volume in the critical lane on each
street at peak hour.
6. A djust the cycle length (sum of all greens and yellows) to the next-higher 5-sec­
ond interval and redistribute extra green time.
7. Com pute percentage values for all intervals.

Example ? \ „ ,
V H f V'
Time an isolated signal with pedestrian indications at the intersection of Pine and Oak:
Pine is 56 ft wide, Oak is 40 ft wide. During the peak hour, the critical lane volumes are 350
and 250 veh/hr and approach speeds are 40 and 25 mph (58.7 and 36.7 ft/see) for Pine and
Oak, respectively (see Pigure 8-E2). ^
O O Ju v w e, k irtaACvivv yaA& ^ ( o ( f

26 Y( 29.5 55
Pine St. 3*-5 S -T -5
vehicles

vnt 2—(\ A-- ---- i 't

•V"*
A 0 W -2§ FDW 2 6 / DW 55 W
Pine St.
O M pk pedestrians %6

5 R
0 RU 31 G 50 Y 53 55 R
Oak St.
/9 5 A
vehicles
, Ct /'-N ‘ :•/
....................... --^
0 DW 3( W 39 FDW 50 DW 55 DW
Oak St. 3! j H 5
pedestrians

H
Figure 8-E2 Phases for Homburger and Kell’s Method.
H,
_________
__________1
Solution - h v i... r
.v p-h.x f HP
1. Select yellow change intervals. rL^-X •
Pine: 3.5 sec Oak: 3 sec /$ ,'> y V ^ V > > 'U r
2. Calculate nondilemma clearance times.
*r '~ 1 5ec a? W --/ a
r V r* , 58.7 40 + 20 „„ "L f
J , — ytPmesl +
20
+ —^ r z ~ = 4.9 sec
58.7
or 5 sec y => 1
g>£L\ UJ>
, 36.7 56_+_20
Q ak ?l + 20 f 5.0 sec or 5 sec £y/ c/-’I*-*) ^ *o“ «
36.7
f i (/- rteV
Calculate all-red clearance intervals.
^ Afterf*ine yellow: 5 - 3.5 = 1.5 sr f CL J = Vl

A fter Oak yellow: 5 - 3 = 2 sec I - A

O-
 29U Intersection C jiu r o l and Design Chap. 8

3. Determine pedestrian clearance tunes. s: - r r " .

; A\ fe
\ K ji'-' Pine (crossing Oak): 40/4 =: iO sec
j^, ■ Oak i cros;>i.jig Pme): 56/4 —
- i4 sec
^ ( FDW (Pine) = 10 3.5 - 6.5 sec
vs/alV*- FDW (Oak) --■= 14 3 - 11 sec

where FDW = Hashing “don’t walk.’’

4. Compute minimum given times (pedestrian clearance - yellow + walking minimum).

pf Pine: '0 - 3,5 +■ 7 — 1^5 sec; use 15-sec minimum ft * Z - 't

tj. 'o r t >o>i S. Compute k*reen tknes (usine Oak as critical minimum';. e s\
t \& \ 1 ------- ■ s u J V s J ^
If / 350(18)/2MJ 2.V2 sec •« 2:? sec ^Pnu Sneet given') *— ^ Q
cycle length'and redistribute extra i>recu in.ie.
. p. q-
Total cycle ~ 25 f 5 + 18 + 5 53 see: use ro see
Extta green time = 55 - 53 — 2 sec; give 1 sec to Pine sec) and 1 sec to
Oak { I #1sec)

1. Compute percentage values foi ail intervals i'or key settings.

W e b s te r's fViethod. W e b ste r u tilized ex ten siv e field o b se rv a tio n s an d c o m ­

p u te r sim ulation to establish an ex cellent p ro ced u re for designing traffic signals. A fu n ­
d a m e n ta l assu m p tio n o f W e b s te r’s w o rk is th a t o f ra n d o m vehicle arrivals. H e
d ev elo p ed the classic e q u a tio n fo r calculating th e average delay p er vehicle on an in te r­
sectio n ap p ro ac h , an d also d eriv e d an e q u a tio n for o b ta :n L ~ IB e p p iiin u i^ cycle tim e
th at pro d u ces the m iniim 'm vehicle d e la /. A u h ough the elem en ts of W e b ste r’s m eth o d
are p re se n te d m this section, rea d ers sh o uld refer to W eb ster arid C o b b e (1962) for fu r­
th e r details.
W eb ster uses term in o lo g y that, n eed s som e ex p lan atio n , so we begin this section
by defin in g such te rm s as “ s a tu ra tio n flo w ” and "lost tim e .” T h is is fo llo w ed by a
d escrip tio n o l average delay an d the o p tim u m cycle tim e. Several exam ples applying
th ese co n cep ts arc in clu d ed .

Saturation F lo w , q s , a n d Lost Time. A stu d y of the d isch arg e o f vehicles

across the sto p bar o f an in te rse c tio n a p p ro a c h in d icates th a t w h en the g re e n p e rio d
begins, vehicles lake som e lim e to sta rt an d accelerate k> norm al ru n n in g sp eed , but
a fte r a few seconds, tire vehicle q u e u e discharges at a m o re-o r-less co n sta n t ra te called
th e su u a a liu n flo w (see F ig u re 8-9). T h e sa tu ra tio n flow is o n e th at w ould be o b ta in e d
if th e re w ere a c o n tin u o u s q u e u e o f veh icles a n d they w ere given 100% g reen tim e. It
is g enerally e x p ic sse d in vehicles p e r h o u r of g re e n tim e. It can be seen from th e figure
th a t the av erag e ra te o f flow is low er d u rin g th e first lew seconds (w hile vehicles are
a c c e le ra tin g to n o rm a l ru n n in g s p e e d ) a n d also d u rin g the a m b e r p e rio d (as so m e
vehicles d ecid e to sto p and o th e rs c o n 'in u ^ to m ove on). It is co n v e n ie n t to rep lace
Sec. 7 Traffic Signals 291

£ : ( %, ; AII'-M
- J^f L.-<< \^Ht.--
Amber
Street A | _
---- KJ Tntergrcen

Amber„
Street B
W///M
Figure. 8-9 S a tu ra tio n F low ,

th e green and a m b e r p e rio d s hv an ^cffecVtVf g rp c n ” p ?iv W .)d u rin g w hich th e flow is

assum ed to tak e p lace at th e .saturation ra te , co m b in ed w ith a “ lost'" tim e d u rin g which
no flow tak es place. T his is a useful co n cep t b ecau se cap acity is then d ire ctly p ro p o r ­
tional to effective g reen tim e . Tn g rap h ical term s, this m ean s rep lacin g th e curve in th e
figure by a re c ta n g le o f o qual area w 'here th e height of th e rectan g le is eq u al to the
av erag e s a tu ra tio n flow an d its b a se is th e e ffectiv e g reen tim e. T h e d ifferen ce
b etw e en th e effectiv e c rre n ;«nie an d the co m b in ed green and a m b e r p eriods is lo s t
time.
S atu ratio n flo"' and lost tim e can be m easu red on the road directly, «nd a m eth o d
for do in g this is as follows: O b serv e o n e lane, o r th e e n tire ap p ro ach , as applicable,
using a sto p w atch to m e asu re th e n u m b e r o f vehicles crossing th e stop b ar, from the
beg in n in g o f th e green p h ase until th e fro n t b u m p e r of th e nth vehicle crosses the stop
line d ir in g e a c 1: s u c :::s iv e 0.1 -m inute interval of green and am ber. T his p ro c e d u re is
best illu stra te d by an exam ple.

Rxunipl

An intersect ion apnrm cb rent rolled bv a fixed-time, sienal v'vis server! for 15 saturated
intervals, each of 0.1. rnrm te. provicfjng Ib t following ‘s.
 292 Intersection Control and Design Chap. 8
c?/
1/>

T im e (m ir ) (in te rv a ls) 0 0.1 0.'.2 0.3 0.4 0.5

N u m b e r o f s a tu ra te d
in te rv a ls 15 15 15 ' 15 15
N u m b e r o t v eh icles
cro ssin g sto p b a r 22 3 ^ 35 30 28 26
jl z
. D isc h a rg e p e r 0.1 m in /" 1.46 /5 ~ * 2 .3 3 2.00 1.86 1.73
X . .C»....

Total duration of ittlthe last saturated intervals = 50 sec

Total number of vehicles crossing stop line = 13
13
Discharge per 0.1 min during last saturated interval X 1.56 vehicles
50/60 10
\ Estimate the iost time. \/_
S o lu tio n

N During the first and last sat urated intervals, there is a Joss of capacity, because vehi-
^ , p cles are accelerating from a stationary position at the beginning of the green period and
; decelerating during the amber period. The flow during the remainder of the observed peri-
\ h t f s ot*s rePreseuts the maximum discharge possible.
' " 3 v>A’ '
2.33 + 2.00 + 1.86 + 1.73
o Saturation flow

C-I W
vvv --- 1.98 vehicles per 0.1 min
U
\\ /•? # y 6 o x ( 0 - 1188 veh/hr
V
The lust time at the beginning and end of the green period may be calculated with refer­
ence to Figure 8-E3. The num ber of vehicles represented by the rectangle efij is equal to
the number of vehicles represented by the original bars of the histogram. Also, the number
of vehicles represented by the area dghk is also equal to the num ber of vehicles repre^-
sf»n‘ed by the IburO.l-min periods of saturated flows between d ancTA:. Therefore^
Area ubdc = area e fg d

and A rea h ijk = area n m lk

Therefore,
ed X 1.98 == 1.46 X 0.1
L- e
=> ed = 0.074 min •£>
and ce — 0.1 - 0.074 = 0.026 min
Similarly,
kj X 1.98 = 1.56 X 0.1
=> k j = 0 ,0 7 9

and ji = 0.1 - 0.079 —.-0-021 min

Sec. 7 Traffic Signals 293

Therefore,
ce + jl = 0.026 + 0.021 = 0.047 min = 2.82 sec
that is, the lost time during the green phase = 2.82 seconds.

fo fa ■ j - &(<'■

3.00
C

e
i a ";
— ------- --------------- —---- SzlJ
f 2.00 r - i •< . i
9
I • |
JC
a>a n
I b m
V.
xf
§* 1.00 '■ /
J<oC
0 C\.
f f ;
■ x ' X s-
d k
0.00- ■' -ZJUL.
0 @ 0.1 0.2 0.3 0.4 0.5 Qj
Time (min)

F ig u re 8 -E 3 O b s e rv e d D isc h a rg e A c ro ss S to p L ine.

Discussion
In this type of experimentation, it is important to note the following points:
1. If the flow on the approach is not saturated, observation?; should be discontinued
until the flow once again reaches saturation level.
2. The distribution of trucks, buses, and other types of vehicles along with turning
movements is also important and should be rrcordpd.
3. The effect of large vehicles and motorcycles is accounted for (in British practice) by
making the iollowing conversions [in passenger-car units (PCUs)]:
1 passenger car or light commercial vehicle -- 1.00 PCI)
1 heavy or medium commercial vehicle = 1.75 PCU
1 bus =(225 PCU
1 motorcycle = 0.33 PCU
1 bicycle =■- 0.20 PCU
1 left-turning vehicle = 1.75 straight-ahead vehicles
4. RigM-tur* "chicles do not affect flow.

^ Average Delay and Optimum Cycle Time. W e b ste r s classic e q u a tio n is

X" /"\l/3
, c (i - or 0.651 yv-~ )' x %tSB
‘ "" 2(1 - Ox) 2q (1 - x)
I
where
A — average delay per vehicle on ar approach (see)
C = cycle time (sec)
294 Intersection Control and Design Chap. 8

8= proportion of cycle length that is effectively green (g /C ) for that phase

q = flow rate (veh/sec)
-,vKy= saturation flow (veh/sec) = 1800 veh/hr oY
x - degree of saturation s *^
= ratio of actual flow to maximum flow through approach
= q/Os
To enable the delay to be estim ated m ore easily, Eq. 9 can be w ritten as
0

CA —— D
Q
A , 13, and D have been calculated and tabulated by W ebster for easy application.
W ebster’s m odel com putes the approxim ate cycle length that minimizes total
intersection delay as well as the effective green time for each approach, by differenti­
ating the equation for the overall delay with respect to the cycle time. The result is

____ 1-5L + 5 _____ ____ J . 5 L + 5__ j , .

° 1-0 - (Yj + Y 2 + ■■■ + Y j 1 0 _ £ Yir

where '■? . ..........l~.L .1

j. ■k ' ■’
-C0 -- optimum cycle length (sec)
( L = total lost time per cycle, generally taken as ihc iam ol the total yellow and
all-red clearance per cycle (sec)
Y = observed volum e/saturation flow, for tHe critical approach in each phase
'w
N ote that this optim um cycle is the one th at gives the highest ratio of flow to satu­
ration flow.
I ne distribution of green time to each phase is proportional to the critical lane
volumes on each phase. For a two-phase intersection the net green time is
G =■■ Q - A 1 - A 2 - nL ^__ -_____3 # #
where
G — net green time (sec)
, C0 -- optimum cycle length (sec)
Al- yellow change interval in phase 1 (sec)
A-y -- yellow change interval in phase 2 (sec)
n = num ber of phases (n = 2)
L = lost time per phase (sec)
W ebster’s equation can be utilized to select the cycle length for a given intersec­
tion. It yields cycle lengths (C0) that are usually shorter than those that are obtained by
other methods, which indicates that cycle lengths should be as short as possible. W eb­
ster also concluded th: ,~’,'y is n o t 1:g n if\ antly increased by cycie-length variation in
n .. a ,
 Sec. 7 T ra ffic S ignals < ''v / < / \ A y 295

the range 0.75C0 to 1.5 Q . However, his equation is sensitive to errors in estim ates of
vehicle flow rate and saturation flow.

Design of a Fixed-Time Signal for a Two-Phase Installation (No Exclusive

Turning Lanes or P!— es). The following steps are indicated:
N1
1. For each approach, calculate or m easure the saturation flow,
2. For each approach, count the peak-hour volume in mixed traffic with known p er­
cent composition and percent turns; divide by the peak-hour factor; convert to
design volume in straight-through passenger cars per hour, using the rough coef-
^ ficients shown in Example 3.
i ~y <;
7 Fo” each rpprc~~1i, calculate the q/f% ratio. For each street, choose the larger
qlgfcfor design. If q a n d ^ arc per lane rather than per approach, choose the crit­
ical lane q /fy;ratio, following the same procedure.
4. Calculate the two intergreen periods.
5. Calculate the minimum green indications based on pedestrian requirem ents.
6. Calculate optimum•
7. Split the available effective green time betw een the two phases.
.
8 Check if ti.e mkiiuium green indications required by pedestrians are satisfied. If
not, adjust upward.
9. Develop a table of all signal indication lengths, according to the following rules:
(a) T h e m inim um length o f any g reen indication is 15 seconds.
(b) The cycle lengtfi'sTiould B~e“a<3justed to the nearest higher length divisible by
5 (if C < 90 seconds) or 10 (if C > 90 seconds). Redistribute the extra green
as described earlier.
(c> All intervals shonld be integer percentage points of the cvcle 1ength.

Example 4
A simple four-leg intersection needs a fixed-time signal. The critical flows in the N-S and
E -W directions are 600 and 400 veh/hr. Saturation flow is 1800 veh/hr and the lost time per
phase is observed to be 5.2 seconds. Determine the cycle length and distribution of green
(see Figure 8-E4).

0 23 30 Iso
Phase NS R G R
......

30 46 50
Phase EW

F ig u re 8 -E 4 P h a se D i g r a m s fo r E x a m p le 4,
296 Intersection Controi and Design Chap. 8

Sulutiou
Assume an amber period of 4 seconds.

Y =
obse:' ved flow
saturation flow
600 .................. . 400
1800 = 0333 ^ = I 800 = a222
S Y; = 0.3,33 + 0.222 = 0.555
1.5L + 5.0 1.5(2 X 5.2) + 5 __ 20.6
C°~ ~1" - 1 Y: I - 0.555 ” 0.445 46,3 SCC
Use C, -■ 50 sec
G — Cu —A ; —A 2 - n!
= 50 - 4 - 4 - 2(5.2) - 31.6 sec ^
X critical lane volumes = 600 -f 400 = 1000

600(31.6) , t
° ' - i “ '“ iuoo"” ■ 1 ! , '%

400(31.6)
' 1000... = 1~ 6‘*
Phase,v_i- = green + yellow + lost time
= 18.96 + 4 + 5.2 = 28.16 sec ~ 30 sec
Phase£_w = 12 64 4 -f 5.2 = 21.84 sec « 20 sec
Cycle length = 30 + 20 =• 50 sec

Example 5
A n intersection, as shown in Figure 8-E5, is to be provided with: five-phase operation,
7-second minimum green time (G), 4-second amber, and minimum flashing “don’t walk”
(FDW) time equal to the time required to walk from curb to the center of the farthest lane
at a rate of 4 ft/sec.
Assumptions: 20-ft average vehicle length; 2.1-second average saturation flow head­
way; random arrival pattern. Use the Australian formula shown in what follows.
Solution
1. Webster’s optimum-cycle formula:

1.5L + 5
C° ~ ~ \ - Y
where
C0 -- optii....... cieto ru i aiiizv .lelay (sec)
y -- volume/saturation flow for the critical approach in each phase
Sec. 7 Traffic Signals 297

500
(5% T.) N
800
500 11
100 100 (10% T.)
300
200

1200 ) 900
100
(10% T .)'
500
(5% T.)
»A) 100
Sec. 7 T ra ffic S ignals 301

Phase 5: (173/1104) X 95 - 15 sec (G + A )

Phase 6: (530/1104) X 95 = 46 sec (G + A )
Phase 4: (401 /1104) X 95 = 34 sec (G + A )
Based on a O^-^cond cycle and a G = 27 seconds (34 —7 losses) for phase 4, the equiva­
lent car factor for the south approach of 2.71 PCU results in a critical volume for lane D of
417 veh/hr. Adjusting the calculations for this volume results in an optimum cycle of
98 seconds. The final splits are
Phase 1: 25 see Phase 5:15 sec
Phase 2: 37 «ec Phase 6: 47 ?r'r
Phase 4: 36 sec Phase 4: 36 sec
P^sed r r these splits and cycle length, a capacity analysis will yield level-of-service
values that can be. used to evaluate alternative phase strategies. The Y value determined in
the “W ebster” analysis is related to volume/capacity ratio and level of service. A Y value
of 0.70 or below indicates a level of service of C or better. A Y value greater than 0.70 indi­
cates a level of service of D or worse. Anytime that Y exceeds 0.70, a serious capacity
problem exists and measures should be taken to increase capacity.
Optimum cycle length also relates to level of service, although the correlation is not
nearly as accurate as that provided by Y. As a general rule, if the optimum cycle calcula­
tion yields a cycle length in excess of 75 seconds for two critical phases, 100 seconds for
three critical phases or 140 seconds for four critical phases, the intersection warrants more
study and measures to increase capacity should be considered.
Assuming the example intersection as the base condition, other alternatives have
been investigated in an effort to increase capacity. Each case is identified as follows:
Case I: example intersection, five-phase operation
Case2: example intersection, five-phase operation, add right-turn lane on west approach
Case 3: example intersection, sir phase sph'f street operation
Case 4: example intersection, six-phase split side street operation, add right-tui l
lane on west approach
Case 5: example intersection, six-phase split side street operation, add right-turn lane
on west approach, widen side streets to provide left-turn lane and two through lanes
Case 6: example intersection, eight-phase operation, add right-turn lane on west
approach, widen cross street to provide left-turn lane and two through lanes
Results of each investigation are as follows:

C o n tro lle r C ritical M in im u m O p tim u m

C a se p h a se s p h a se s cycle (sec) cycle (sec) Y

1 5 3 54 98 0.65
2 5 3 57 89 0.62
3 6 4 77 164 0.73
4 6 4 80 143 0.69
5 6 4 83 114 0,61
6 8 A 111 0.60
302 Intersection Control and Design Chap. 8

Cases 1,2,5, and 6 should be considered for improving iruersectiou efficiency. Cases
3 and 4 should be rejected as feasible alternatives.

J*ignataro's Method. A n alternative m ethod of designing a four-approach

two-phase signal cycle follows, as suggested by Pignataro (1973). By referring to defin­
itions regarding PH F and headway, let
N x = m ajor street critical lane flow, the num ber of vehicles in a single lane,
A/j being the largest lane flow of the two mnjor street approaches *
9
N 2 = minor street critical lane flow, similar to N l
C = cycle ler^iii ($ec)
S { = approxim ate average headway entering intersection among
S2 - corresponding average headway among /V2
Y\ = vehicle clearance interval for N x (sec)
Y2 = vehicle clearance interval for N 2 (sec)
K = num ber of signal cycles for a 15-minute period
Then the total time required to pass all vehicles through the intersection during the
15-minute period is

and total time required for clearance interval is

K ( Y 1 + ,y 2) (14)
Therefore,
T -i- K ( Y t + Y2/ < 15 min X 60 - 900 sec (15)
or, in the limiting condition,

(16)

But

_ 9 0 0
(17)
C
By substituting Eqs. 13 and 17 in Eq. 16, we have

Yi + y > (18)
min
1 - ( Nl S 1 + Ar2S2)/3600(PH F)
or in the general case when there are three or more phases,
„ _________ 2 Y ;
___

( 19)
Cm, " l ' ZNi S; / 3600(PHF)
Sec. 7 Traffic Signals 303

Equation 18 can he further modified to reflect an exclusive pedestrian phase as follows:

Yx + Y2 + P + Yp
r = -------- 1------±------------- E------ no)
" 1 - 0 .0 0 0 3 3 3 ( ^ 5 ! 4- N 2S z) k '

where P is the length of the pedestrian “walk” interval (sec), and Y p is the length of the
pedestrian clearance interval (sec).
Assuming that the average tim e spacing (headway) betw een vehicles on the
major and minor approaches is about the same, the allocation of green time is made
proportional to the critical lane volumes. Therefore,
G, + G2 = C - (Yi + Y 2) (21)
and
n , _ c - ( r , + y 2)
G ^ - ^ 2 [ C - ( Y l + Y2)] (22 )
1 + N JN X

and
,2 - C - (Yj + Y 2) - Gj
Go (23)
For intersections where and S 2 (the headways on m ajor and minor streets) are
not assumed to be t^~ same, the green time should be allocated in the ratio of the prod­
uct of critical lane flows and its corresponding headways:

(24)

Example 6
An isolated intersection of Ftye and O^k Streets needs a simple two-phase signal. The fol­
lowing dala arc ,,. ailable: »
PJbe Street: 56 ft wide, critical lane volume —300 veh/hr, approach speed 40 mph
Ozfc Street: 40 ft wide, critical lane volume = 225 veh/hr. approach speed 25 mph
10% trucks and 15% left turns, PtIF = 0.85
Pedestrian walking speed = 4 ft/sec L
Solution ? % .?cc
*9
See Figure 8-E6. 1 f
Gi R,
Pine

23 28 50 !0 'j
I SV

r2 g2 v2

Oak

28 45 50

8 -E 6 F ie n n ta ro ’s M e th o d .
 3H r
304 /K i Intersection C ontrol and Design Chap. 8

1. Determine adjusted critical lane volumes: truck factor 1.5, LT 1.6.

LrV, 'T" ‘1
N * ^ Ejffe Street: 0.15(1.6) (300) + 0.10(1.5) (300) + 0.75(1) (300) = 342 veli/hr
a/. ■=£> Oirffe Street: 0.15(1.6) (225) + 0.10(1.5) (225) + 0 75(1) (225) = 257 veh/hr
' C •'» 7' C.
2. Determine vehicle clearance intervals (nondilemma limes).

v W + I
Y- t+ —
la v

w h e re *

l ™ 1 see
y 4 0 n :r- 10 i t/s e c "
„ v I ^ 2u ft
Y Street (crossing Q^fe): Y ~ 1 + 58.7/20 + (40 + 20)/58.7 = 4.96 sec (say, 5.0 sec)
y/ O ^ . Street (crossing Pine): Y — 1 + 336.7/20
6 + (5o + 20)/36.7 = 4.91 sec (say, 5.0 sec)
3. Determine pedestrian crossing limes:
t, /
/ i Vv / f,
Crossing Qqfc: 7 + (4 0 /4 )- 5 — 12 sec = C, - ?
Crossing Pjpe: 7 +(56/4)— 5 = 16 sec = G2

where 7 sec is the assumed pedestrian startup time.

4. Determine approximate cycle length: Based on vehicle consideration (PHF = 0.85
and .Vj - s2 - £.5 sec):

c - r '-+ y 2 ..........
1 - (A',5, -t- iV252)/3600(PHF)
4.96 + 4.91
1 - (342 X 2.5 + 257 X 2.5)/(3o0<) X 0.85) * scc
Based on pedestrian requirements:

Qim ~ G] + G2 + V’j + Y2 = 12 + 16 + 4.96 + 4.91 = 38 sec

(Tiiis is the controlling value.) ; .1' ^ y ”
5. Split the cycle:

__ C - (V'j + Y^) 38 •- (4.% 4.9 S)

' .' i ' T N M ~ ~ ...... 1 + 257/342 .. " CC

2 1 + N x/N 2 1 + 342/257
6. Determine fbf rvcle lenr"b
C - 16 + 16 4 5 + 5 = 42 sec (say, 45 sec, giving an extra 2 sec to J^fe and
i see to O^k)
- -
f .—^
L=r,

id 't~ 17 -t 5 + = 4:? sec

G - ■---/?
7. Check the peak 15-nunute requirement for ea.:h phase.
 & r 9t.| \x.
Sec. 7 Traffic Signals ^ ..*»» u 305

REQ! "RED '■'"'EEIM TIME PROVIDED GREEN TIME

/'/■ S- 342 900 f) r c

7 >/TnjYr'. ' x 2.5 251 see .- — X 18 = 360 sec
4(0.85) ■ 45 — " - c.

n,
. 0 _. X 2.5 = 189 sec
—— X 17 = 340 sec
u- 4(0.85) .
. 45,. , 7-
c JVu' ■ ; s'- (! ,.
7.6 Traffic-Actuated Signals (Mitric, 1975} t V ' ' ; :YV'/ Vfc"

The success of fixed-cycle signals depends on-the, degree- to which #ctirar traffic vol-*>
umes agree with design flfyy rates built into the design process discussed before. A sin-
gle cycle time and a single set of-greci. indications Would be difficult-tb* justify for an
entire 24-hour period, in view of the fluctuations in hourly traffic volumes. This prob­
lem can be partially solved by designing several’fixed-cycle schemes;for an intersec­
tion, each based on a traffic flow prevalent during a given part of the day. , ,
Traffic-actuated signals represent an extension of the idea described before:
M ake the cycle time and green splits responsive to changes in traffic flow,, down to the
level of microchanges. That is. the length of green indication could be varied by the
num ber of vehicle; arrivals. , ^
The simplest type of traffic-actuated installation has a detector located at a dis­
tance A ahead of the stop line at an intersection approach^ and a controller sensitive to
signals sent by the detector. A t the beginning of a green phase, the maximum num ber ‘
of vehicles caught between the stop line and the detector (including a vehicle stopped
exactly over the detector) is given by

_ A_ (25)
T + s0

where I is the average vehicle length, ^nd 5, is the d: bet ween nearest points of
consecutive stopped vehicles. The maximum indicated green for this approach is

Gmin = fsd + ( t V T + 1 ) T (26)

J + s0

w here

qs = saturation flow (pcu/hr)

startup delay

The controller’s memory starts at “blank” in each green phase. If within G min sec­
onds it does not receive any “calls” from the detector, signifying additional vehicle
arrivals, the green may be switched to another phase. For each call, however, ttie con­
troller will extend the green time by a fixed-time interval, called the unit extension of
green, counted from the moment of the call. The unit extension must be long enough to
allow a vehicle to cover the distance from the detcctor to the stop line moving at a
given approach speed, v:
306 Intersection Control and Design Chap. 8

h = A /v ' (27)
where h is the unit extension of gre6n (also called a vehicle interval). N ote that for
every call that the controller receives, the green time is landed by h svconds, meaT
sured from the ftiomeht when that vehicle detected. In other words, there is no accu­
m ulation Unused gteeri time A hefef6re,rsubeessive vehicles may be no m ore than h
seconds apart to keep the green time at thfeir approach. Because interval h is used to
determ ine w hether gr«: — .1111 be eA-::]nd'<L or term inated for a given approach, it is
sometimes called the critical gap (or critical headway). It is im portant to note that the
critical gap and the unit extension of green are one and the same.
If vehicles keep following each other at headways shorter or equal to the unit
extension, the green time for this approach will be term inated only if it reaches a pre­
set maximum based on maximum individual delay to vehicles on other approaches or
on some other delay-based criterion. If all approaches are saturated, traffic-actuated
installation operates in a fixed-cycle m anner. The unit extension is clearly the most
im portant param eter in the design of traffic-actuated signals; and the question is: How
long should it be? ■ ■
Safety considerations dictate that a detector be located no closer than the safe stop­
ping distance ahead o f the stop line. This impH6sa:mininiiirn length qf unit extension:

= (28)

A bove this minimum, the length of the unit extension depends on traffic arrival
rates. A s with fixed-time signals, it can be selected by referring to average delay or
num ber of stops. Intuiuvciy, shorter c*nit extensions should be used when arrival rates
are high, and vice versa. E xperim ental work shows that the optim al (average^delay-
minimizing) unit extension ranges from 2 to 8 seconds, as a function of average traffic
flow per phase. This is illustrated in Figure 8-10.
It follows that simple traffic-actuated signals suffer from some of the same weak­
nesses as those of fixed-time signals. They will work well if the actual traffic flow
matches the flow assumed when the unit extension of green was selected. A n obvious
direction of im provem ent is to design traffic-actuated signals with variable unit exten­
sion, varying with the length of green and/or with the changes in flow rate. Such
iiriprovements are, in fact, operational.

^ TOrfffofSigwals (Mitrte, 1975). For isolated inter­

sections (0.5 mile Or nlore distant frt>m adjacent signiri«controlled intersection) that
experience sharp fluctuations in flow during the day, the following procedure could
be adopted.

1. For a given intersection approach, choose the headway, h, between vehicles that
, is just enough to ’ „! he greeu ndic^aon. This is usually in the interval between
2 and 5 seconds. This headway is equal to the unit extension of green per vehicle.
2. Calculate the distance, A , upstream from the stop line at which the detector will
be located, by the formula
Sec. 7 Traffic Signals 307

C
O
oc>
X
03

c
D
E
□
E
a
O

Average tra ffic flo w per phase (veh/hr)

F ip ir c 8-10 A v e ra g e T raffic Flov/ p e r P h a se versu s O p tim u m Uni* E x ten sio n .

A = v x h (ft) (29)
where v is the approach speed of vehicles (ft/sec), and h is the headway from step
1 (sec). This i r ^ n s that h seconds after the nth actuation the light will turn amber
just as the nth vehicle is at the stop line, unless the (n + l)th actuation occurred.
3. Calculate the length of the minimum green’that should be long enough to pass all
the vehicles that might, be accumulated between the stop line and the detector,
plus one vehicle.

A
G mm ~ lsd + (30)
\l + Sr
where
t. - startup delay (sec)
A — distance between the stop line and detector (ft)
I = vehicle length (ft)
S0 = distance between rear end and front of the consecutive stopped vehicles (ft)
qs = saturation flow (FCU/hr)
4. Calculate the elements of the lost time: intergreen (am ber, all red) and amber
effectively used as green, in the m anner previously described.
5. R epeat the calculation for each approach; all components of the minimum cycle
are now ready.
298 Intersection Control and Design Chap. 8

L ~ total lost time/cycle, generally taken as the sum of the total yellow and all-red
clearance per cycle (sec)
2. Austni ian left-turn factor formula:

EL , _ _ _ (12)

GnS-Q GnS - Q G
where
0

E L = factor for converting unprotected LT vehicles to equivalent passenger car units

S — saturation flow (veh/hr/ln)
~ headvv^/
C -■ cycle length (sec)
n = number of opposing lanes
G — effective green time (sec), de ined as the green plus amber time for a phase,
minus startup and terminatior losses
Q - exposing straight-through and right-turning traffic (PCU/ph)
/ -- £ function of the opposing flow (<2 ) that varies as follows:

Q /

0 1.00
200 0.81
400 0.65
600 0.54
800 * 0.45

3. Minur.um-cycle calculation: Assuming considerable pedestrian activity, the mini-

m um eycle length will be controlled 6y the m i n i n ^ i tinifij set for phase 1 + 2 + 4
or by phase 5 -t- 6 + 4.
Phase 1 = 7 sec, 7 + 4 sec, A = 11 sec
Phase 2 = 7 sec, W + 52/4 FDW = 2U sec
Phase 4 = 7 sec, W + M /4 FDW = 23 sec
^ Total =- 54 sec
The |£Hal time for phases 5, 6, and 4 i; the same.
4. Optimum cycle: Convert street volun es to equivalent passenger-car units (PCUs).
Ideally, conversion factors should be (developed from field measurements. However,
for this example, the following factor have been assumed:
Strai- ht-through passenger car: 1.0 P ^U
f LT p tssenger car during protected p^ase: 1.1 PCUs
RT passenger car with turning radiu 25 ft: 1.1 PCUs
RT passenger car with turning radiu' < 25 it: 1.3 PCUs
Truck or buses: 1.5 PCUs
LT passenger cars during a permissive phase are equal to the value obtained from
the Australian L uuuu.. For this jxam^ie, S =? 1714, a 90-second cycle, and a 30-second
Sec. 7 Traffic Signals 299

effective green time for phase 4 have been assumed, resulting in a factor of 2.38 PCUs for
the north approach anc^2.6] PCUs for the south approach. Applying these factors to the
traffic volumes results in the following equivalent passenger-car volumes:

A p p ro a c h L e ft S tra ig h t R ig h t
------------------ 7 — ------------ ^ v X ’->?)
W e st ' " ' " - ^ 231 945 116 ^
Eabi. 173 525 173
N o rth 244 308 113
S o u th . 401 256^ 133 ((fyl
•h-Z—

Applying the approach volumes to the intersection lane configuration results in the fol­
lowing hourly volumes:

ip p ro a J i L an e P h a se M ovem ent V o lu m e

W est A 1 1 231
W est 'B 6 6 (945 + 11 6 )/2 = 5301
W est C 6 6 (945 + 1 1 6 )/2 = 531 )
S o u th D 4 4 401
S o u th E 4 4 (256 + 133) - 389
E ast F 5 5 173
E ast G 2 (525 + i7 3 ) /2 = 3497,
E ast H 2 2 (525 + 1 7 3 )/2 = 349 >
N o rth I 4 8 244
N o rth .T 4 8 3 ^ 3 , 421

The cycle length will be controlled by the critical lane volume for the pha^s combination of
(1 + 2 + 4) or (5 + 6 + 4). *
Phase 1 lane A = 231
Phase 2 lane G or H = 349 •. u. 4,!n{ , U
*vl L* ;l ^ f\ / ^y v>
Phase 4 lane D = 401^ 7
Total = 981 1^
Phase 5 lane F =173
Phase 6 lane B or C = 530
Phase 4 lane D = 401
Total \ =1104 '■ , /
Phase (5 + 6 + 4) will control the cycle length. The ratio of critical flows to saturation flow
can now be calculated:
total critical flows 1104
y = ------------------- -_ = ------ ~~ 0.644
saturation flow 1214
Lost time per cycle can now be calculated. Lost time consists of the time lost during queue
startup and time lost during phase *-rm ir f’on. Startup l&ses vary with the queue length
as follows:
300 Intersection Control and Design Chap. 8

Q u e u e (P C U s) S tratu j) losses (see)

1 1.7
2 2.7
<1 3 3.3
J 4 3.6
5 3.7
6 3.7

A s s u m in g a 90-second cycle, splits p r o p o r t i o n a l to critical la n e vo lu m e s, and a uniform

arrival rate, the average queue length can be calculated. Average arrival rate (g):
Phase 5; 173 veh/ln/hr = 0.048 veh/ln/sec —x
Phase 6: 530 veh/ln/hr = 0.147 veh/ln/sec
Phase 4: 401 veh/ln/hr = 0.111 veh/ln/sec
Red time (R): /< 7 , , , ..
— c ' ' k: f0
g t Pht.se 5: [(1104 - 173)/11Q4] X 90 = 76 sec
Phase 6: [(1104 - 530)/1104] X 90 = 47 sec
Phase 4: [(1104 - 401)/1104] X 90 = 57 sec
Average queue length (g X R):
Phase 5: 0.048 X 76 = 3.6 vehicles (£«•
Phase 6: 0.147 X 47 —6.9 vehicles » %
Phase 4: 0.111 X 57 = 6.3 vehicle^
Phase-term ination losses include the time consumed while the last vehicle in the
queue travels from the stop bar to a point safely beyond the traveled way of the next con­
flicting movement. It is assumed that these vehicles are traveling at the approach speed,
except for the LT phase vehicles (phase 5), where 15 mph has been assumed. .
Phase 5: (50 + 20)/22 = 3.2 sec W ^
Phase 6: (60 + 20)/44 = 1.8 sec
Phase 4: (75 + 20)/29 = 3.3 sec
Combined startup and termination losses are as follows:
Phase 5: 3.6 + 3.2 = 6.8 sec
Phase 6: 3 .7 + 1 .8 = 5.5 sec
Phase 4:3.7 + 3.3 = 7.0 sec >
Total = 19.3 sec -- L fa lr i '■ ^
The optimum cycle to minimize delay

1.5L...+—
( —— 5 ----------------------
(1.5 x 19.3) + 5—
J 1 Y 1 - 0.644
Splits are as follows:
 *
Sec. 7 Traffic Signals 309

increase in rear-end collisions. Further, the installation of signals may not only increase
overall delay, but also reduce intersection capacity. Consequently, it is of the utmost
im portance that the consideration of a signal installation and the selection of equip­
ment be preceded by a thorough studv of t^ ffic and <-~^^ay conditions by an engineer
experienced and trained in this field. This engineer should recognize th at a signal
should be installed only if the net effect, balancing benefits versus drawbacks, is to the
public’s advantage.

7.7 Warrant Criteria for Signal Control

Because of the complexity in deciding when and where a signal should be installed, it is
apparent that some system of establishing the need for a signal installation at a partic­
ular location is necessary. Such a system has been established using a common denom ­
inator known as sig n a l w arrants. The eight w arrants described briefly here are
discussed in detail in the M a n u a l on U niform T raffic C ontrol Devices fo r Streets and
H ig h w a ys (M UTCD: F1TWA, 1988). These should be considered as a guide to the
determ ination of the need tor traffic control signals rather than absolute criteria, and
their use tem pered with professional judgm ent based on experience and consideration
of all related factors. For example, such factors as physical roadway features, age of
pedestrians, or effect of adjacent signalized intersections may modify a decision based
solely "n the w a rra r^

Warrant 1: Minimum Vehicular Volume. The minimum vehicular warrant is

satisfied when, for each of any 8 hours of an average day, the traffic volumes given in
Table 8-1 exist on the major street and on the higher-volume m inor-street approach to
the intersection. These m ajor-street and minor-street volumes are for the same 8 hours.
D uring each hour, the higher approach volume on the m inor street is considered,
regardless of its direction

Warrant 2: Interruption of Continuous Traffic. The interruption of continu­

ous traffic w arrant is satisfied when for each of any 8 hours of an average day, the traf­
fic volumes given in Table 8-2 exist on the m ajor street and on the higher-volume

TABLE 8-1 M IN IM U M VEHICULAR VOLUME W ARRANT 1

V ehicle p e r h o u r
V e h icles p e r h o u r on h ig h e r-v o lu m e
N u m b e r of lan es to r m o v in g ' o n m a jo r s tre e t m in o r-stre e t a p p ro a c h
tra ffic on cach a p p ro a c h (to ta l o f b o th a p p ro a c h e s) (o n e H irection only)

M a jo r stre e t M in o r stre e t U rb a n R u ra l U rb a n R u ra l

500 350 150 105

2 o r m o re 600 420 150 105
2 o r m o re 2 o r m o re 600 420 200 140
1 2 o r m o re 500 350 200 140

.V o w rrF H W A , 1088.
310 Intersection C on trol and Design Chap. 8

TABLE 8-2 M IN IM U M VEHICULAR VOLUMES FOR W ARRANT 2

\ ch icle pe* h o u r
V e h icle s p e r h o u r o n h ig h e r-v o lu m e
N u m b e r o f laiies fo r m o v in g o n m a jo r s tre e t m in o r-s tre e t a p p ro a c h
traffic o n ea c h a p p ro a c h (to ta l o f b o th a p p ro a c h e s) (o n e d ire c tio n only)

M a jo r s tre e t M in o r S tre e t U rb a n R u ral U rb an R u ra l

1 1 750 525 75 53
2 o r m o re 1 900 630 75 53
2 o r m o re 2 o r m o re 900 530 ^00 70
1 2 o r m o re 750 525 to o 70

Source: FHWA., 19S8.

minor-street, approach to the intersection and where the signal installation will not seri­
ously disrupt progressive traffic flow. These m ajor-street and m inor-street volumes are
for the same 8 hours. During each hour, the higher volume on the minor street is con­
sidered, regardless of its direction. A reduced volume, similar to that described under
w arrant 1, can be used in place of those shown in Table 8-2 uu highcr-speeu roads or in
smaller communities.

Warrant 3: Minimum Pedestrian Volume. This w arrant combines pedestrian

volumes and available gaps in vehicular traffic at either intersection or midblock loca­
tions, The w arrant is satisfied when, on an average day, there are 100 or m ore pedes­
trians for each of any four hours, or 190 or m ore pedestrians in one hour crossing the
m ajor street. These values may be reduced by as much as §0% where the predom inant
pedestrian crossing speed is below 3.5 ft per second. In addition, during the period
when the pedestrian count is satisfied, there shall be less than 60 gaps (of adequate
length for pedestrians to cross) per hour. This requirem ent applies to each direction
separately on a divided street with a median wide enough for pedestrians to wait. A
signal may uot be warranted if the proposed site is included in coordinated traffic sig­
nal system that allows less than 60 gaps per ’ >ur as such a signal may provide fewer but
longer gaps for pedestrians to cross safely.
This w arrant applies only to locations where the nearest signalized intersection is
m ore than 300 ft away-and the proposed new installation will not unduly affect p ro ­
gressive flow'. A t midblock locations, curb parking should be prohibited at least 100 ft
in advance of the crosswalk and for at least 20 ft beyond. Street lighting should be con­
sidered if the midblock location is used at night. Normal pedestrian signal indications
shall be installed under this w arrant and should utilize a traffic actuated signal con­
troller and pedestrian deiectors (push buttons). It shall operate within a coordinated
scheme if the signal is in coordinated system.

Warrant 4: S ch o o l Crossings. The fourth w arrant recognizes the unique

problems related to children crossing a m ajor street on the way to and from school,
particularly near the school, and may be considered as a special case of the pedestrian
w arrant. A traffic control signal may be w arranted at an established school crossing
Sec. 7 Traffic Signals 311

when a traffic engineering study of the frequency and adequacy of gaps in the vehicu­
lar traffic stream, as related to the num ber and size of groups of schoolchildren at the
school crossing, shows that the num ber of adequate gaps in the traffic stream during
the period when the children are using the crossing is less than the num ber of minutes
in the same period.

Warrant 5: Progressive Movement. The progressive m o v e m e n t w arrant

relates to the desirability of holding traffic in compact platoons. This warrant: applies
when the adjacent signals are so far apart that they do not provide the necessary degree
of platooning and speed control on a one-way or two-way street. In addition, for a two-
way street, the adjacent signals should constitute a progressive signal system. A ccord­
ing to this w arrant, the installation of a signal should not be considered where the
resultant signal spacing would be less than 1000 ft (305 m).

Warrant 6: Accident Experience. T he accident experience w arrant is sat­

isfied when

1. An adequate trial of less restrictive rem edies with satisfactory observance and
enforcem ent has failed to reduce the accident frequency.
2. Five or more reported accidents of types susceptible to correction by traffic signal
control have occurred within a 12-month period, each accident involving personal
injury or property damage apparently exceeding the anplicabie requirements for
a reportable accident.
3. T here exists a volume of vehicular traffic not less than 80% of the requirements
specified in warrants 1 or 2.
4. The signal installation will not seriously disrupt progressive traffic flow.

Warrant 7; Systems. This w arrant recognizes that traffic signal coordination

can be two-dimensional. That is, progression along an important cross street can be as
im portant as that along what would normally be called the major street. Both streets
must be given equal consideration as major routes. The systems warrant is applicable
when two or m ore major routes meet at a common intersection and the total existing or
immediately projected entering volume is at least 1000 vehicles during the peak hour of
a typical weekday or each of any 5 hours of a Saturday or Sunday.

Warrant 8: Combination of Warrants. In exceptional cases, signals may occa­

sionally be justified when no single w arrant is satisfied, but where warrants 1 and 2 are
satisfied to the extent of 80% or m ore of the stated values. The 80% requirem ent
under this w arrant is applied against 70% requirem ent tnr rura1 areas, uiaking it 56%
of that applicable to urban areas. A dequate trial of other rem edial m easures which
cause less delay and inconvenience to traffic should precede installation of signals
under this warrant.

Warrant 9: Four-Hour Volume. This w arrant uses 4-hour volumes instead of

8-hour volumes used in warrants 1, 2 and 8. The warrant defines curves representing
312 Intersection Control and Design Chap. 8

vehicles per hour on the m ajor street (both directions) and on the higher-volume
m inor-street approach (one direction only). This w arrant is satisfied (for urban loca­
tions) when, for each of any 4 high hours of an average day, the plotted points fall
above the specified curve for the existing combination of approach lanes.
The requirem ents are lower when the 85th percentile speed of m ajor-street traffic
exceeds 40 mph, or when the intersection l-ies within a built-up area of an isolated com­
munity with a population under 10,000. A separate series of curves is specified for
4-hour volume requirem ents of rural locations. A reference should be made to the Man­
ual of U niform Control Devices (M UTCD) (FH W A, 1988) for a complete discussion.

Warrant 10: Peak-Hour Delay. For this w arrant to be satisfied, all of the fol­
lowing condiiions should be met:

1. The total delay experienced by the traffic, during the peak hour, on a side street
controlled by a ^'T'OP sjgn e -,r.als o- exceeds five vehicle-hours for a two-lane
approach and four vehicle-hours for a one-lane approach.
2. The volume on the side-street approaches equals or exceeds 150 vph for a two-
lane approach or 100 vph for a one-lane approach.
3. The total entering volume serviced during the peak hour equals or exceeds
800 vph for intersections with four or m ore approaches or 650 vph for intersec­
tions with three approaches.

Warrant 11: Peak-Hour Volume. This is intended for intersections where

m inor-street traffic experiences undue delay or hazard in entering or crossing the main
street. The w arrant is based on a critical combination of main-street and cross-street
volumes during any four consecutive 15-min periods during the peak hour of an aver­
age day. A series of curves is specified for traffic volume on major street (total for both
directions) and the higher-volume m inor street (vehicles per hour for one direction
only) separately for urban and rural locations. A signal is w arranted when m easured
volumes fall above the specified curve for a given combination of approach lanes. A
reference should be made to M UTCD for a complete dL^ussion.

7.8 Coot dination of Traffic Signals

Some form of signal coordination is necessary on major streets having a series of inter­
secting streets in order that vehicles flow without having to stop at every intersection.
Signals can be coordinated in several ways, but the three most common techniques are
the simultaneous system, the alternative system, and the flexible progressive system.
1. Simultaneous system. In this technique, all signals along the coordinat
length of the street display the same aspect to the same traffic stream at the same time.
This system reduces capacity and encourages the tendency to travel at excessive speeds
so as to pass as many signals as possible. It is best suited where the city blocks are short.
W here turning traffic is light, it may have advantages for pedestrian movement. Some
local control can be introduced using vehicle actuation, but a m aster controller keeps
ail the local controllers in step and imposes a common cycle time.
 Sec. 7 Traffic Signals 313

2. Alternative system. H ere, alternative signals or groups of signals give opposite

indications at the same time, which means that if a vehicle travels the distance between
intersections in half the cyclc time, it need not ston. The cycle time must be the same
for all signals, so that the speed of pi egression is constant.
3. Progressive system. Two types of progressive systems are used. In the simple
progressive system, the various signal faces controlling a given street give green indica­
tions in accordance with a time schedule to allow continuous operation of a platoon of
vehicles t~ flov nt a planned speed. In the flexible progressive system, the intervals at
any signal may be adjusted independently to the traffic requirem ents and in which the
green indications at separate signals may be started independently at the instant that
will give the m axim um efficiency. A m aster controller keeps the local controllers,
which may be fixed-time or vehicle-actuated, in step.

Balanced Two-Way Signal Progression. A balanced directional progression

is often desired on streets during off-peak hours or during the entire day. In such
cases, a simple method of obtaining the widest equal two way “through” bands is
desired. Purdy’s (1967) method is one such technique that can be applied to a new or
existing series of intersections. This quick-and-easy method is best described through
an example.

Example 7
Design a signal progression that will give the maximum equal two-way “through'’ band­
width give’-' +hat speed 30 mph = 44.1 ft/sec; cycle = 60 sec; distances in feet: A B = 800,
A C = 1800, A D = 2300, A E = 2.900, A F = 4100; arterial green: A = 60%, B = 50%,
C = 70%, D = 50%, E = 60%, F = 70%.
Solution
>
The most convenient way to assemble the data and calculations is in the form of a
table.

In te rse c tio n

L ine A B n D E F
\
O 1 0 800 1800 2300 2900 4100
2 0% 30% 68% 87% V10% 155%
^ 3 0 -2 0 + 18 -1 3 J-10 +5
' 4 60% 50% 70% 50% 60% 70%
30/ -3 0 25/ -2 5 35/ -3 5 25/ -2 5 30/ -3 0 35/ -3 5
i q / -3 0 + 5 / -4 5 53/ -1 7 12/ -3 8 40/ -2 0 40/ -3 0
* 6
7 30/ -3 0 + 5 / -4 5 + 3 / -6 7 12/ -3 8 40/ -2 0 40/ -3 0
—tr —
> '\ .
r - / Line 1: cumulative distance in feet to each intersection from base intersection A.
’ x, r ^ Line 2: speed = 30 mph = 44.1 ft/sec. Convert distance in feet to travel time in sec-
,^ V> onds and in terms of cycles (as a percentage). 1 cycle = 60 sec = 2646 ft/cycle (e.g.,
y* ~ tjv Y at B, 800/2646 = 30%). f
 Intersection Control and Design Chap. 8

Line 3: the algebraic numerical difference between the nearest multiple of 50% and
the percentage appearing on line 2 (e.g., x - 50%, x - 100%, x - 150%).
Line 4: pen. ^ green u. iven
Line 5: each green is placed in its initial reference position with respect to the com­
mon datum line.
Line 6: add lines 5 and 3 algebraically (num erator and denominator separately). The
smallest “plus” = 5 and the smallest “minus” = 17. The through band in both direc­
tions = 5 + 17 = 22% of cycle.
Line 7: it may be necessary to adjust values in the num erator and denominator by
adding + 5 0 /+ 5 0 or - 5 0 / - 5 0 to any values in line 6 to obtain a wider bandwidth.
For example, C can be (+ 5 3 / —17) + ( —5 0 /—50) = + 3 / —67. In this case, the band­
width works out to 3 + 20 = 23% of the cycle.
From Figure 8-11 and the preceding procedure, steps for drawing the tim e-speed
diagram follow:
1. Use half of green time at A as a benchmarl... The band at A will have a portion below
the benchm ark equal to the smallest “ph.'s” and a portion above the benchmark
equal to the smallest “minus” of the cycle ime.
.
2 Draw a band in both directions according to the speed limit of the vehicles.
3. Find the offset time for each intersection by subtracting the line 7 values from line 2
values. If line z is more than i00, suotract 100 anil u^r* t he remainder as the line 2 value.
4. Draw the signal intervals based on these offsets for each intersection.
5. N ote that offsets are given in percentages, and then converted to seconds (e.g.,
- 3 0 X 60 = 18 secs).

Bench
mark

A B C D E F
+30/—30 +5/—45 + 3/-67 +12/—38 + 40/-20 + 40/-30
—30(—18 sec) 25(15 sec) 65(39 sec) 75(45 sec) 8 seo) ^ 15(9 sec)

Figure 8-11 P u rd y ’s M e th o d .
 SUMMARY

Traffic control devices arc an im portant part of an engineer's k i t of tools for dealing
with the efficient m ovem ent of vehicles and pedestrians along streets and through
intersections. In this chapter, much emphasis has been given to the general topic of
traffic signals, because they are much m ore difficult to design and operate than are
other methods of traffic control. First, the basis for understanding the components of
traffic signals was provided, followed by a theoretical examination of signal systems. A
num ber of practical field methods of designing isolated intersection signals were given,
followed by a brief description of designing traffic-actuated signals.
Eleven w arrants specified by the M anual on Uniform Traffic Control Devices
(FHW A, 1988) were explained briefly, followed bv a short description of a method to
coordinate signals on an arterial in progression.
For those inclined to obtain a m ore com prehensive knowledge of intersection
design and control would do well to refer to D rew (1968), FH W A (1985), and JH K &
Associates (1980).

REFERE NCES

A m e r ic a n A s s o c ia t io n o f St a t e H ig h w a y JHK & A s s o c i a t e s (1980). Design o f Urban

and T r a n s p o r t a t io n O fficla ls ( A A S H T O ) Streets. Technology Sharing Report 80-204,
(1990). A Policy on Geometric Design o f High­ U.S. Departm ent of Transportation, Wash­
ways and Streets, A A S H T O , Washington, DC. ington, DC.
A m e r ic a n A sso c ia t io n o f St a t e H ig h w a y M it r ic , S. (1975). Transportation Engineering
and T r a n s p o r t a t i o n O f f i c i a l s (AASHTO) notes, Ohio State University, Columbus.
(1994). A Policy on Geometric Design o f High­ P i g n a t a r o . L. J. (1973). Tra ffic Engineering
ways and Streets (Metric), AASHTO, Wash­ Theory and Practice, Prcnrice-Hall, Engle­
ington, DC. wood Cliffs. NJ.
R. ( 1 9 6 8 ) . Traffic Flow The­
D re w , D o n a ld P l in e , J. S. (Ed.) (1992). Traffic Engineering
ory and Control, McGraw-Hill, N e^7YorV H a...'' k, 4th Ed., Prentice Hall, Engle­
(FHWA)
F e d e r a l H ig h w a y A d m in ist r a t io n wood Cliffs, NJ.
(1985). Traffic Control Systems Handbook, P u r d y , R. J. (1967). Balanced Two-way Signal
U.S. D epartm ent of Transportation, Wash­ Progression, Traffic Engineering, Institute
ington, DC. of Traffic Engineers, Washington, D C
(FHWA)
F e d e r a l H ig h w a y A d m in is t r a t io n S a l t e r , R. J. (1974). Highway Traffic Analysis
(1988). Manual on Uniform Traffic Control and Design, Addison-Wesley, Reading. MA.
Devices for Streets and Highways, U.S. T r a n spo r t a t io n R e s e a r c h B o a r d (T R B )
Department of Transportation, Washington, (1985). Intersection Channelization Design
DC. Guide, N CH RP Report 279, National
G r e e n sh ie l d s, B. D.. D. S h a p i r o , and E . L. Research Council, Washington, DC.
E r i c k s o n ( 1 9 4 7 ) . Traffic Performance at T r a n s p o r t a t i o n R e s e a r c h B o a r d (TRB)
Urban Street Intersections, Technical Report (1994). Highway Capacity Manual, Special
1, Bureau of H i g h w a y Traffic, Yale Univer­ Report 209, National Research Council.
sity, New Haven. CT. Washington. DC.
H om burger, W. S. e t a l . ( 1 9 9 6 ) . Fundamen­ W e b s t e r , F. V., and B. M, C o b b e (1962). Traf­
tals o f Traffic Engineering, 14th E d . , Univer­ fic Signals, Road Research Technical Paper
sity of California. Berkeley. 56. Her Majesty's Statirnery Office, London.