5: INTERSECTION DESIGN AND CONTROL

5.1 INTRODUCTION

Intersections play an important role in any network system. They are the points where
traffic flow converges and where direction of travel changes. Intersections may be categorized
according to shape, type of structure, and type of operation.

Shape
This refers to the configuration of the intersection and would depend largely on the
number of legs. Carefully planned and properly designed road networks often lead to
intersections with simpler shape having lesser number of legs. Some of the most commonly
observed intersection configuration is the following:

a. Three-leg: T or Y

b. Four-leg: normal crossing, oblique, or skewed/staggered

c. Multileg: intersections with more than four legs

5: INTERSECTION DESIGN AND CONTROL

d. Rotary or roundabout

Type of structure
Most of the intersections are either designed as at-grade intersection or grade separation
such as flyovers or interchanges. Almost all intersections are initially designed at-grade and are
planned to be grade-separated in the future to cope with high traffic volume. The type of grade
separation depends largely on the extend of improvement it would provide in terms of easing
congestion or reducing traffic accidents.

Type of operation
Rules and regulations applied to a given intersection depend largely on the type of control
which is in operation at that intersection. The main objective is to simplify traffic flow. This is
often achieved by reducing the number of conflicts of vehicles.
An intersection operates as unchannelized or channelized, and unsignalized or signalized.
Channelization often leads to simplified movements of vehicles as it leads drivers to one conflict
at a time. On the other hand, signalization greatly crossing conflicts at the intersection area.

5.2 BASIC INETRSECTION DESIGN PRINCIPLES

When designing or improving an intersection, it is very important to follow some basic

principles to avoid costly mistakes that lead to lack of the intersection’s capacity or to its being
accident-prone.
a. The maximum number of legs should be four. It has been shown that the number of
conflicts increases exponentially as the number of intersection legs increases.
5: INTERSECTION DESIGN AND CONTROL

b. Staggered intersections should be avoided. Due to proximity of the two side roads,
inadequate length of storage lane for left vehicles often causes blockage of the through
traffic.
c. Main traffic flow should be near straight as possible. Sharp turns, such as left and right
turns, cause unnecessary reduction in traffic speeds.

d. Roads should not intersect at a small angle. Oblique intersections pose potential hazards
and cause high severity of accidents due to the almost head-to-head collision of vehicles.
An angle of 60 degrees must be considered as the minimum.

e. Two intersections should be as far as possible from each other. In addition to the reason
cited in (2), adequate weaving sections may not be provided. As a rule of thumb, the
distance between the two intersections must be

distance = design speed (kph) x number of lanes x 2

For instance, if the prevailing speed is 30 kph and three are four lanes in one direction,
the distance between the two intersections must be at least 30 x 4 x 2 = 240 m.
5: INTERSECTION DESIGN AND CONTROL

5.3 INTERSECTION DESIGN ELEMENTS

Turning geometry
The principal purpose of an intersection is to provide change in the direction of travel. As
a vehicle approaches an intersection, the driver has to decide whether to go straight or to turn to
left or right. For turning movements, a number of turning geometries may be considered, the
most direct of which is highly preferred. Configurations of turning geometries are shown below:

a. Direct

b. Semidirect

c. Indirect
5: INTERSECTION DESIGN AND CONTROL

These types of turning movements may be found as elements of the interchanges shown
below:

The different elements of an intersection are shown in figure 5.1. The adequacy in design
of each element must be carefully checked, considering traffic flow and availability of right of
way.
5: INTERSECTION DESIGN AND CONTROL

For the left turn storage bay, the recommended length is

= 2 x no. of left turners in one cycle x spacing in queue.
The spacing in queue may range from 6.0 m to 7.0 m, depending on the type of vehicles
using the intersection. A factor of 2 is used to take into account randomness of traffic flow.
The number of lanes for through, right, and left turn vehicles would depend on traffic
volume, saturation flow rates, a simple circular curve may be adequate for the design of the
intersection corners and the turning roadway. However, simple curve may not be enough for
large intersections. The most common type of geometry for these intersections is the three-
centered curve as shown in figure 5.2.

Figure 5.2
Three-centered curve

5.4 METHODS OF CONTROL OF INTERSECTIONS

Conflicts often occur at intersections. The more the number of legs an intersection has,
the more the number of conflicts it has. Figure 5.3 shows the number of conflicts of three-leg and
four-leg intersections.
5: INTERSECTION DESIGN AND CONTROL

Figure 5.3
Conflict diagrams for three-leg and four leg intersections

Conflicts may be classified as merging, diverging, or crossing conflicts. Table 5.1 gives a
summary of these conflicts. The total number increases exponentially with the number of legs of
the intersection. Most of the problems at intersections, like congestion and accidents, are caused
by crossing conflicts.

Table 5.1
Types and number of conflicts

Depending on the traffic volume using the intersection and the severity of conflicts,
intersection control may fall under any of the following categories:
a. Unsignalized
b. Signalized
c. Grade separation
5: INTERSECTION DESIGN AND CONTROL

5.4.1 Unsignalized Intersection

When two minor roads intersect, traffic may be assumed to arrive at any approach of the
intersection in random fashion. In this case, signals may not be necessary but some of control
must be employed to ensure safety. The simplest form of control is to have no control at all. At
least, there are no visible reminders such as signs or markings posted in the vicinity of the
intersection. There is actually a rule for minor intersections where there is no clearly defined
priority. The so-called right-of-way rule governs. This rule states that when two vehicles arrive at
the same time at the intersection, the vehicle on the right has the priority. This rule must be
understood by all drivers and it does not depend on the size of the vehicle! There is no such thing
as “right of weight”.
In the case of intersections where priority is clearly defined – a major road crossing a
minor road, for instance – the traffic on the minor road has to depend on available gaps at the
major road to be able to cross or turn. For this kind of intersections, reasonable controls for
unsignalized intersections may be done through the use of traffic signs such as YIELD or STOP
sign. The use of YIELD sign is to establish or strengthen legal superiority of one traffic flow
over another where most traffic on all legs is able to proceed through without a full stop. With
the use of STOP sign, two-way control, or four-way stop control may be implemented. In a two-
way stop control, the major approaches without STOP signs have complete priority over the
minor approaches with STOP signs. In the four-way stop control, all approaches have equal
priority, Studies indicate that four-way stop control works the best advantage of traffic when
flows on the two cross streets are approximately equal. In a simulation study of four-way stop
control by Wright, it has been found out that the percentages of right, left and through
movements has a little effect on delay and queue lengths. This may be due to the fact that four-
way stop control offers flexibility, i.e., the driver is given the opportunity to use his own
judgement in deciding what movements are possible at a given instant.

5.4.2 Roundabout or Rotary

This is another form of unsignalized intersection that works best when the number of
turning vehicles is almost equal to the number of through vehicles. In a roundabout, all vehicles
move counter clockwise in one-way fashion. Movement is characterized by merging at very
5: INTERSECTION DESIGN AND CONTROL

small angles, weaving, and diverging when leaving the intersection. When traffic signals’ cost
proved to be costly, roundabout is a good alternative.

Figure 5.4
Roundabout (all vehicles move in one-way fashion)
(Photo taken by author.)

5.4.3 U-Turn Slots

The U-turn scheme has been implemented at several intersections within Metro Manila
for a number of years now. Operation at U-turn slots is very similar to that of a roundabout
except that it favors a particular road, which is assumed to be the major road.

Figure 5.5
A U-turn slot in Quezon City
(Photo taken by author.)
5: INTERSECTION DESIGN AND CONTROL

5.4.4 Signalized Intersection

The control of traffic intersections by traffic signals is done by separation in time, i.e.,
conflicts between opposing or merging streams are prevented by giving the right of way to a
given direction. Section 5.7 provides an introduction to traffic signal control.
Consider a four-leg intersection with all movements allowed. Figure 5.6 shows the
conflict diagrams for the different types of intersections discussed previously.

Figure 5.6
Conflict diagrams for different types of control for a four-leg intersections
5: INTERSECTION DESIGN AND CONTROL

Table 5.2 shows a comparison in terms of number of conflicts and the applicable type of
control or regulation for these four types of intersection.

Table 5.2
Comparison of types of control for a four-leg intersection

5.4.5 Grade Separation

Grade separation eliminates the problematic crossing conflicts of the different
movements of vehicles. Flyover/overpass, underpass, or full-blown interchanges are some of the
examples of grade separation. Some of these grade separations or interchanges are shown in
figure 5.7. They provide the safest and most efficient method of control but definitely the most
costly, and sometimes unaesthetic.
5: INTERSECTION DESIGN AND CONTROL

Figure 5.7
Grade separation or interchanges

5.5 ANALYSIS OF UNSIGNALIZED INTERSECTIONS

A large number of intersections all over the country are still without traffic signals. It is
therefore necessary to have a means of analysis of the performance of this type of intersection so
as to find appropriate measures to minimized congestion and reduce the occurrence of traffic
accidents.
The method presented in this section was the basis of the method discussed in the US
Highway Capacity Manual with some modifications. In our case, it may be better to introduce
the original method. As new researches on the topic are conducted, modifications may be done to
suit local conditions.
The method calculates the maximum flow in any given minor road traffic stream. It is
them compared with the existing traffic flow to estimate the reserve capacity. The probable delay
and level of service are determined based on this reserve capacity.
5: INTERSECTION DESIGN AND CONTROL

Consider the four-leg intersection below. Minor movements are movements coming from
the side road or minor road. In addition, the left turn movements from the major road are also
m=considered minor. But in terms of hierarchy, these left turn movements have higher priority
than the movements coming from the minor road.

The method requires that the traffic movements be dealt with in the following order:
a. Right turns into the major road
b. Left turns off the major road
c. Traffic crossing the major road
d. Left turns into the major road

5.5.1 Structure of Major Road Traffic

Table 5.3 shows the different major road traffic streams, which have to be considered
when analyzing a particular minor road flow.
Note that for the traffic turning right from the major road (m1), which actually does not
have direct conflict with the minor flow being considered, half of its volume is considered as
inhibiting effect. If this inhibiting effect is shown to be absent through actual observation of the
condition, M1 may be totally omitted.
5: INTERSECTION DESIGN AND CONTROL

Table 5.3
Major road traffic streams

5.5.2 Critical Gap

The source of capacity of the minor road flows is the available gaps between major road
flows. The critical gap used to describe the minimum gaps needed by drivers of minor road
vehicles. Values of critical gaps are given in table 5.4 for different vehicle maneuvres, speed
limits, and highway type.

5.5.3 Capacity
In the estimation of capacity of the minor road flow, the basic capacity is initially
determined. Based on the major road flows given by Mh, and values of critical gap tg, the value
of the basic capacity Mno is read from the graph shown in figure 5.8.
5: INTERSECTION DESIGN AND CONTROL

Table 5.4
Critical gap tg for passenger cars, sec.

Figure 5.8
Basic capacity of minor road flow
5: INTERSECTION DESIGN AND CONTROL

The basic capacity is the maximum minor road flow, assuming that the following
conditions are true:
a. The traffic on the major road Mh does not block the major road.
b. A turning lane is provided for the exclusive use of the minor road traffic stream.

If these conditions are not met, correction factors have to be implied based on the
following considerations:
a. Congestion on the major road
If a traffic stream turning off the major road becomes congested and thus
interferes with the minor road traffic, then the basic capacity is reduced. A p-value is
obtained from figure 5.9 and is applied as a reduction factor. The p-value defines the
probability that this minor road traffic stream remains unaffected. The ration of the actual
flow and the capacity of the same flow is initially used to determine the p-value.

Figue 5.9
Reducation factor due to possible congestion of movement of major road
5: INTERSECTION DESIGN AND CONTROL

b. Shared lanes
Shared lanes on the minor road approach are lanes in which two or more
movements are confine in the same lane. When corners of the intersection have a large
turning radius, the effect of shared lanes may be minimal or may be ignored since minor
vehicles can stop side by side at the edge of the major road.
The capacity of the share lane can be determined using the equation

(5.1)
where Mn - capacity of all streams using the shared lane
Mna, Mnb, Mnc - capacity of individual streams
a, b, b - contribution of the individual streams to the total volume using the
shared lane.

5.5.4 Passenger Car Equivalents

Since the calculated capacity is in pcu/hr, the existing flow has to be converted to the
same unit. Table 5.5 may be used for this purpose.

Table 5.5
PCU values of different vehicle types
5: INTERSECTION DESIGN AND CONTROL

5.5.5 Reserve Capacity

The difference between the existing flow and capacity is termed as reserve capacity. The
magnitude of delay and level of service are directly related to this reserve capacity. Table 5.6 is
used for assessment purposes. Overall evaluation or generalization may have to be done after
assessing each minor traffic flow.

Table 5.6
Reserve capacity

Example 5.1
Consider the channelized T-intersection with the minor road controlled by YIELD sign.
There is no prevailing speed limit. Evaluate the performance of the unsignalized intersection.
5: INTERSECTION DESIGN AND CONTROL

Solution:
Following the logical order of analysis as suggested:
a. Right turn into major road

b. Left turn off major road

5: INTERSECTION DESIGN AND CONTROL

c. Left turn into major road

Overall evaluation: The left turn traffic into major road experiences very long delay.
There are no problems with the other minor flow. The left turn off major road is not likely to
block the through traffic.

Example 5.2
Evaluate the unsignalized intersection with the two minor road approaches controlled by
STOP signs. The traffic volumes shown are in vehicles per hour (vph). There is no prevailing
speed limit.
5: INTERSECTION DESIGN AND CONTROL

Solution:
Due to symmetry of traffic volume data, only the analysis of the minor road approach C
will be shown. The same condition is expected in approach D.
a. Right turn into major road

b. Left turn off major road

5: INTERSECTION DESIGN AND CONTROL

c. Crossing major road

d. Left turn into major road

5: INTERSECTION DESIGN AND CONTROL

Overall evaluation: the intersection requires improvement. Traffic signals and geometric
improvement (channelization) may be necessary.

5.6 ANALYSIS AND DESIGN OF ROUNDABOUTS OR ROTONDAS

The elements of the roundabout are shown in figure 5.10. Roundabout with large central
island are known to operate better as they provide ample weaving sections that normally dictate
the capacity of the intersection.

Figure 5.10
Some design elements of a roundabout
5: INTERSECTION DESIGN AND CONTROL

For the analysis and design of roundabouts, British practice will be discussed in this
section (Salter 1976(. The focus is the capacity of the weaving section, which is given by the
equation

(5.2)
where w – the width of the weaving section in meters
e – the average width of entries to the weaving section in meters;

l – the length of the weaving section

These variables are defined in figure 5.11, which shows a typical weaving section

Figure 5.11
Weaving section
5: INTERSECTION DESIGN AND CONTROL

The capacity formula is valid under the following conditions, although there is no reason
to believe that the formula does not hold if any variable lies a little outside the values given:

The design of roundabouts is usually done by trial and error. Weaving width and length
are assumed and then checked for adequacy of capacity. As roundabouts have a tendency to lock
when overloaded, it is important that they have adequate reserve capacity.

Example 5.3
Consider the Y-junction with the given traffic volume to be designed as a roundabout.

Table 5.7
Types and number of conflicts
5: INTERSECTION DESIGN AND CONTROL

It is usually convenient to draw the intersection diagram as follows:

With the traffic volume superimposed, section AB is the weaving section with a critical
volume of 1,300 vehicles per hour.
Assuming the following:

The capacity of AB section is calculated:

5: INTERSECTION DESIGN AND CONTROL

The computing for the reserve capacity:

The other sections (AC and BC) may be designed with the same parameters, this
providing large reserve capacity.

5.7 TRAFFIC SIGNAL CONTROL

For intersections with relatively low traffic volumes typically found in rural and suburban
areas, traffic may be assumed to arrive at the intersection in random fashion. Likewise, the same
randomness of arrivals may be observed where spacing between intersections is of considerable
distance. In these situations, signals may not be necessary and the intersection has to be
controlled by traffic signs as discussed in the previous chapter.
The control of traffic intersections by traffic signals is done by separation in time, i.e.,
conflicts between opposing or merging streams are prevented by giving the right of way to a
given direction. This separation in time is termed as signal phasing. And it can be seen that the
more the number of conflicts, the more the number of phasing needed. However, in some cases
as in a major street-minor street intersection, or in approaches where left-turning volume is
relatively low, a separate phasing for the minor flow may not be necessary. The number of
phases employed at any intersection must be kept to a minimum, compatible with safety because
with every phase added, there is a corresponding additional loss of green time, which eventually
leads to increased intersection delay.

5.7.1 Types of Signals

Nowadays, many types of traffic signals are being used to control traffic flow. Depending
on the distance and influence of one or more signals over the other, the intersections may be
classified as isolated, coordinated or area controlled.
5: INTERSECTION DESIGN AND CONTROL

Isolated
A particular intersection may be considered isolated if arrivals of vehicles at its approach
are random. Consider the traffic flow at the two signalized intersections A and B are shown in
figure 5.12. Supposing a queue formed in A proceeds toward intersection B. If the queue has
dispersed and vehicles arrive at random at B, then B may be considered as an isolated
intersection. Naturally, this will happen only when distance between the two intersections is far
or when traffic volume is relatively low. In highly built up areas such as the city or town centers,
it is unlikely to have this kind of intersection.

Figure 5.12
Two intersection distant apart (Isolated intersections may be controlled either pre-time or
actuated.)

Pre-time signal
Pre-time or fixed time signals work best when there is not much fluctuation in traffic
flow. The pre-time signal controller may be of a single or multiprogram type of controller. The
single-program controller makes use of one set of signal parameters to control traffic flow
throughout the day or during the period when the signal is in operation. On the other hand, the
multiprogram type makes use of a number of sets parameters. This offers greater flexibility and
may be able to cope with the fluctuating demand within the day (figure 5.13). Nowadays, a
controller with dive or more programs is not uncommon.

Figure 5.13
Programming per-time signal control
5: INTERSECTION DESIGN AND CONTROL

The Manual on Uniform Traffic Control Devices for Streets and Highways (MUTCD)
cites the advantages of pre-time control:
a. Consistent starting time and duration of intervals of pre-time control facilities
coordination with adjacent traffic signals especially when they are on two or more
intersecting streets or in a grid system. This coordination may permit progressive
movement and a degree of speed control through a system of several well-spaced traffic
signals. Pre-time control provides more precise coordination that allows maximum
efficiency in the operation of two or more very closely spaced intersections operating
under the capacity conditions, when the timing relationship between intersections is
critical.
b. Pre-time controllers are not dependent for proper operation on the movement of
approaching vehicles past detectors. Thus the operation of the controller is not adversely
affected by conditions preventing normal movement past a detector such as stopped
vehicle or construction work within the area.
c. Pre-time control may be acceptable than traffic-actuated control in areas where large and
fairly consistent pedestrians volume are present, and where confusion may occur as to the
operation of pedestrian push buttons.
d. Generally the installed cost of pre-time equipment is less than that of traffic-actuated
equipment, and the former is simpler and more easily maintained.

Traffic-actuated
For isolated intersections where randomness of arrivals is expected, traffic actuated type
of signal control is appropriate (figure 5.14). In this system, detectors are located only on the
approaches of the minor road. With this set up, continuous green time may be given to the major
road traffic flow. Right of way is given to the minor road only when demand is detected. In case
of two roads where no clear priority is given to any direction, the fully actuated type of control is
sometimes utilized. In this scheme, all approaches are provided with detectors.
When the interval of time between vehicles crossing the detector becomes greater than
the last unit extension period, the right of way is transferred to the other directions. When there is
heavy volume of traffic in a particular direction, successive demand for extension period may
5: INTERSECTION DESIGN AND CONTROL

happen and this will lead to continuous green indication. Hence, a present maximum green
period is necessary to limit the length of the green period. It can be expected that when traffic
demand is heavy in all approaches, green period may always reach the maximum resulting in an
almost fixed-time operation.

Figure 5.14
Types of traffic-actuated signals

In a general, a minimum green period or shortest period of right of way is given to any
phase. This minimum green period is long enough to clear the vehicles waiting between location
of the detector and the stop line, and it may be extended for as long as continuous demand for the
right of way exists. Such extension periods are individually and not cumulatively set. Figure 5.15
shows how traffic-actuated control works.
5: INTERSECTION DESIGN AND CONTROL

Figure 5.15
Green time extensions for traffic-actuated control

THE MUTCD gives the special advantages of traffic-actuated control, which include the
following:
a. Traffic-actuated control may provide maximum efficiency at intersections where
fluctuations in traffic cannot be anticipated and programmed for with pre-time control.
b. Traffic-actuated control may provide maximum efficiency at complex intersections
where one or more movements are sporadic or subject to variation in volume.
c. Traffic-actuated control will usually provide maximum efficiency at intersections of a
major street and a minor street vehicular or pedestrian traffic, and also by restricting such
interruptions to the minimum time required.
d. Traffic-actuated control may provide maximum efficiency at intersections unfavourably
located within progressive pre-time systems, where interruptions of major street traffic
are undesirable and must be held to a minimum in frequency and duration.
e. Traffic-actuated control may provide the advantages of continuous stop-and-go operation
without unnecessary delay to traffic on the major street, whereas isolated pre-time signals
are sometimes switched to flashing operation during periods of light traffic.
f. Traffic-actuated control is particularly applicable at locations where traffic signal control
is warranted for only brief periods during the day.
g. Traffic-actuated control tends to reduce any hazard associated with the arbitrary stopping
of vehicles.
5: INTERSECTION DESIGN AND CONTROL

Coordinated
Where randomness can no longer be ensured and there is a need for continuous
movement over an arterial, coordination or synchronization of the timing of the signals in series
is required. A number of methods are used to achieve this. This methods are discussed at the last
section of this chapter.
It is also possible that within an arterial, subgroups consisting of a number of
intersections may be developed (figure 5.16). Each subgroup is treated independently from
another subgroup.

Figure 5.16
Subgrouping of intersections for efficient control

Area traffic method

In urban centers, links (road sections) and nodes (intersections) often form a network
(figure 5.17). Optimization techniques have been developed to cope with such network.
Nowadays, sophisticated software run by powerful computers are sued. Examples of these are
the SCTAS, SCOOT, etc.

Figure 5.17
Road network in urban centers suited for area traffic control
5: INTERSECTION DESIGN AND CONTROL

At present, the Sydney Coordinated Adaptive Traffic System (SCATS), a dynamic

control system is installed in Metro Cebu and some parts of Metro Manila. In Metro Cebu, it was
claimed to have improved the traffic situation in the city. It was also believed that the system has
helped inculcate discipline among the drivers, commuters, and pedestrians.
SCATS and SCOOT (Split Cycle Offset Optimization Techniques) are considered as
third-generation system, in which the signal timings are free to evolve in response to detected
traffic volumes and queues. While demand-responsive systems are more expensive both to install
and to operate compared with fixed-time systems, when operated in the correct environment they
usually achieve benefits that compensate for their capital expenditure within the first few years
of operation. In the SCAT system, small subareas of up to ten intersections share a common
cycle length, which can be altered by p to 6 seconds once per cycle according to the degree of
saturation of that subarea. It can strategically combine certain subareas for varying lengths of
time to improve on overall network performance. It draws its data from stop line detectors. Each
subarea has five offset plans. These internal offsets between adjacent intersections within a
subarea are predetermined as part of the input database. Internal offsets can vary according to the
current cycle time and an input parameter known as the progressive speed factor, which governs
the percentage change in offset. There are also five external offset plans for the purpose of
“marrying” or merging two subareas. These plans are selected by an algorithm that may be used
to select internal offsets. When two adjacent subareas are married, the common cycle time for
the combined area is the larger cycle time of the two separate subareas before linkage.

5.7.2 Data Requirements for Traffic Signal Setting

In planning for phase pattern and for the determination of appropriate timing of the
signals, a number of data are required:
a. Traffic Volume – must include all types of vehicles (including nonmotorized: bicycles,
pedicabs, etc); directional
b. Pedestrians flows – movement of pedestrians in all directions
c. Passenger car unit values (PCU) – PCU equivalent of the different types of vehicles
d. Saturation flow rates – maximum flow rate occurring at the stop line once traffic initially
in queue is given green time indication
5: INTERSECTION DESIGN AND CONTROL

e. Physical characteristics of the road – number of legs; width of approaches; number of

lanes; gradient.
Consider the flow of a single queue of vehicles once signal indication turns to green as
shown in figure 5.18. At the start of green, it can be observed that the lead vehicle will take a
second or two to react. It will be after two or three vehicles crossing the stop line when the traffic
flow will be at maximum rate. This brings about what is called starting loss. In the diagram, the
starting loss is estimated as the time from the actual start of green up to the point where the line
representing the maximum rate crosses the horizontal axis. The maximum flow rate that
corresponds to minimum headway is also known as the saturation flow rate. The value of
saturation flow rate depends on a number of factors: vehicle mix, condition of the pavement,
geometric characteristics of the road (width of the lane, gradient), direction of movement, etc.

5.7.3 Phase
Phasing is a process by which right of way is given to particular movements in a logical
manner with the primary purpose of minimizing the number of conflicts. The number of conflicts
at an intersection depends largely on the geometry of the intersection and turning prohibition. It
may be observed that as the number of legs of the intersection increases, the number of conflicts
increases exponentially.

Figure 5.18
Saturation flow rate determination
5: INTERSECTION DESIGN AND CONTROL

In planning for the appropriate phase pattern for a particular intersection, the major
factors o be considered are the traffic volume and any turning regulation/prohibition. Oftentimes,
it is the left turn movement that gives the problem at an intersection as it complicates and
increases the number of the required phases. Nevertheless, the basic principle is to minimize the
number of phases commensurate with safety.
Given the geometry of the intersection and a set of traffic volume, there are actually
many possibilities of developing phase patterns. The role of the traffic engineer is to find out the
most appropriate one to meet the requirements of meeting the traffic demand and providing
safety to all road users.
Some of the most common phase patterns are shown in figure 5.19:

a. Two-phase system

b. Three-phase system
Figure 5.19
Typical phase patterns

Again, in developing appropriate phase patterns, each phase should not combine any
major conflicting movements. In the two-phase system shown in figure 5.19a, for example, the
through and the right turn traffic may have some conflict. However, this is normally of diverging
type and may not pose a problem as long as the vehicles stay on their respective lanes. Phase 1,
for example, does not allow a left turn movement. This is the situation when this movement is
prohibited in the intersection or when the number of left turn vehicles is so small that it does not
warrant a left turn phase. However, when left turn movement becomes problematic, it is often
5: INTERSECTION DESIGN AND CONTROL

necessary to provide a separate left turn phase as shown in figure 5.19b. Here are some basic
guidelines when planning for the phase patterns.
a. Start with the most problematic/heavy movement. Add movements one at a time and see
to it that they are not in conflict with the previous movements.
b. The number of phases must not exceed the number of legs of the intersection. (Note:
patterns with overlapping movements will be discussed in a separate section. They are
not really counted as phases.)
c. Rearrange the phase patterns so as to follow a logical order.

5.7.4 Lost Times

In the previous section, it has been said that the number of phases must be minimized
commensurate with safety. The major reason for this is that as the number of phases increases,
the amount of lost time also increases. The lost time is composed of the starting loss and the all-
red period. In figure 5.20, it can be observed that for every phase, there is a corresponding
starting loss and all-red period.
The generally accepted value of starting loss is 2 seconds. The amount of all-red periods
depends on the size of the intersection. Together with the yellow or amber period, they compose
the clearance interval or intergreen period. With the yellow or amber having a value of 3
seconds, the all-red period has to be long enough for the last vehicle crossing the road to clear
safely.

Figure 5.20
Lost times consisting of starting loss and all-red period
5: INTERSECTION DESIGN AND CONTROL

5.7.5 Cycle
A cycle is one complete indication of green, yellow and red. Its length is constant for a
pre-time or fixed time signal. It varies for traffic-actuated signals depending on traffic demand.
The fundamental relation between intersection delay and cycle length is shown in figure
5.21.

Figure 5.21
Cycle and delay relation (Co provides the minimum delay.)

It can be seen that using either a long or short cycle will lead to long delay, with very
short cycle giving worse results. The objective of signal timing is to minimized delay. There are
other objective functions – travel speed, fuel consumption, emissions, etc.- but they are all
related to delay. Reducing delay would lead to better speed, fuel savings, and lesser emissions.
The cycle corresponding to least intersection delay is termed as the optimum cycle length.
Webster’s formula is widely used to estimate this cycle length. The equation is given by

(5.3)
where L – total lost time
Y – sum of y-values
The consumption of cycle will be illustrated by an example.

Example 5.4
Consider the traffic volumes and saturation flow rates for the different movement:
5: INTERSECTION DESIGN AND CONTROL

Compute for the optimum cycle time.

(Use yellow = 3 sec; All red = 3 sec; Starting loss = 2 sec)

Solution:
Analyzing the different movements, directions, and traffic volumes, a two-phase signal
may be appropriate.

For all movements, the y-values are computed.

For phase 1, the critical y-value is 0.325 while for phase 2, the critical y-value is 0.517.
This gives
Y-value = 0.325 + 0.517 = 0.842
For this two-phase signal, the loss time L is:
L = 2 x starting loss + 2 x all-red = 2 x 2 + 2 x 2 = 8 sec.
From equation 5.1, the optimum cycle length is:

(The cycle length is normally rounded off to multiples of 5 or 10.)

5: INTERSECTION DESIGN AND CONTROL

5.7.6 Allocation of Green Times

After the cycle length is computed, the total green time has to be allocated to the different
movements or phases. This is normally achieved by dividing it in proportion to the y-values. For
a particular phase, i, for example, its green time is estimated as follows:

(5.4)

Example 5.5
Consider the data in the previous example, allocate the green time to phases 1 and 2.

Solution:
The computed optimum cycle is 110 sec and the total lost time is 8 sec. This results to
total usable or effective green g:

5.7.7 Capacity of Movement or Approach

Given the amount of green time allocated to an approach or movement and its saturation
flow rate, the capacity of such approach or movement can be estimated as follows:

(5.5)
where
s – saturation flow rate
g – effective green
C – cycle length
5: INTERSECTION DESIGN AND CONTROL

Example 5.6
In the previous example, determine the capacity of the south approach.
Solution:
The south approach has a saturation flow rate
s = 1,800 pcu/hr of green time
Its effective green is equal to the effective green of phase 2 (g2 = 63 sec)
Therefore:

5.7.8 Estimation of Delay

The equation for the optimum cycle length was actually derived by minimizing the delay
function obtained by Webster. Webster’s formula for delay consists of three terms. The first term
is the major component of the delay equation. It can be easily obtained by considering the area if
the triangle in figure 5.22. This is actually based on a D/D/1 queuing where both arrival (q) and
departure (s) patterns are assumed regular. The second term based on Allsop was obtained by
assuming an additional queue interposed between the arriving traffic and the signal. The last
term is a correction factor obtained from simulation.

(5.6)
The first term is the dominant term in the formula representing 90-95 percent of the total
average delay. From figure 5.22, this component of delay will be derived.

Figure 5.22
Delay representation per cycle
5: INTERSECTION DESIGN AND CONTROL

Assuming it takes to to dissipate the queue after the start of green:

The total delay per cycle D is equal to the area of the triangle in figure 5.22.

(5.8)
Therefore, the average delay per vehicle per cycle is

(5.9)

Example 5.7
Using equation 5.9, estimate the average delay per vehicle per cycle of movement 3.

Solution:
Movement 3 has a y-value equal to 0.325. Its effective green is 39 sec.
Therefore, its effective red is
r = c – g1 = 100 – 39 = 71 sec
Using equation 5.9, the average delay is
5: INTERSECTION DESIGN AND CONTROL

5.7.9 Timing Diagram

After all the signal parameters have been computed, it is better to produce the timing
diagram where all the numerical values are displayed. This is later applied to the signal
controller. Figure 5.23 shows a sample timing diagram for a two-phase signal system.

Figure 5.23
Timing diagram for a two-phase system

Example 5.8
Draw the timing diagram for the problem in example 5.4

Solution:
The timing diagram is normally drawn using the actual green (not the effective green).
Actual green is computed as follows:
actual green + yellow = effective green + starting loss
(Note: yellow is considered usable and is part of the effective green)
Since yellow = 3 sec and starting loss = 2 sec
actual green = effective green – 1
For phase 1: actual green, G1 = 39 – 1 = 38 sec
For phase 2: actual green, G2 = 63 – 1 = 62 sec
The timing diagram is shown with the different components in seconds.
5: INTERSECTION DESIGN AND CONTROL

5.7.10 Intersection Degree of Congestion

Although the Y-value gives the indication whether to continue with the computation or
not, the degree of congestion X, is often used as the measure of the performance of the
intersection. For a given approach or movement, the degree of congestion of that approach or
movement is the ratio of the demand over capacity.
Let xi be the degree of congestion of a particular movement or phase.
i,
then

(5.10a)
On the other hand, the intersection degree of congestion X is given by

(5.10b)
It can be shown that the degree of the intersection is also the same as the degree of any
movement of approach if the allocation of green is based on the proportion of y-values.

Example 5.9
Compute for the degree of congestion of the intersection described in example 5.4.

Solution:
5: INTERSECTION DESIGN AND CONTROL

With the cycle, Y-value and L computed previously, equation 5.10b can be used to
determine the degree of congestion.

Equation 5.10a can also be used to compute for the intersection degree of congestion since green
times are allocated based on the proportion of y-values.
Consider movement 1 with volume = 930 pcu/hr and saturation flow rate of 1,800 pcu/hr.
Based on y-value of 0.517, its green time was computed as 62.6 sec. (This was rounded off to 63
sec in the previous calculations).
The degree of congestion of movement 1 based on equation 5.10a is

which is equal to the degree of congestion X.

5.7.11 Critical Movement Analysis

When phase patterns become complex, cycle time and green times needed by the
different movements may be determined using the critical movement analysis method. This is the
same method used by Akcelik in his Signalized and Unsignalized Intersection Design and
Research Aid (SIDRA) program and by the US Highway Capacity Manual for the design and
analysis of signalized intersections.
For a given combination of intersection geometry and traffic volume, a finite number of
logical phase patterns may be identified. (Logical phase pattern is defined here as a phase pattern
that is normally accepted in traffic engineering practice in terms of safety and driver
expectation.)
Let m – total number of possible phase patterns
pi – ith phase pattern
The best phase pattern P that is expected to result in minimum delay is

(5.11)
where Ycri is the sum of y-values of critical movements in Pi.
5: INTERSECTION DESIGN AND CONTROL

Figure 5.24 shows the three typical “stages” for vehicles at a T-intersection. The
corresponding critical movement diagram is shown in figure 5.25. The term “phases” is not used
because of the overlapping movements (defined as movements present in more than one stage),
namely: 1, 4, and 5. The cycle time and green time may be dictated by the time requirements of
any of the following sets: [1, 6], [3, 5], [2, 3, 6], or [2, 4]. Each sets consists of movements that
form a complete cycle. The set that gives the highest Y-value forms the critical movements. In
other words:

(5.12)
where n is total number of possible sets.

Figure 5.24
Three stages for a T-intersection

Figure 5.25
Critical movement diagram for the T-intersection
5: INTERSECTION DESIGN AND CONTROL

Example 5.10
Consider the T-intersection with given traffic data:

Compute for the optimum cycle time.

Solution:
Based on the traffic flow, the phase pattern or stages shown below may be considered:

The y-value of each movement is computed.

5: INTERSECTION DESIGN AND CONTROL

The y-values are superimposed in the critical movement diagram

Each of the following sets consists of movements that form a complete cycle:

Therefore, the critical movements are [2, 3, 6] with Y-value = 0.658.

Since there are three major movements involved, 3 starting losses and all-red periods are
expected, giving a total loss time L = 3 x (2+2) = 12 sec.
The optimum cycle is computed using equation 5.4

Allocating green time will be the same as in example 5.5

5.7.12 Treatment of Left Turn Traffic

If left turn traffic is allowed at the intersection, three types of control may be
considered: (1) without phase. (2) with fixed time left turn phase, and (3) with actuated left turn
phase.

Without left turn phase

As a rule, multiple phases should be avoided whenever possible because they generally
require an increase in the overall cycle length. Other options may be considered; left turn bays,
shorter cycle lengths, leading or lagging greens or both, and turn prohibitions. In the case of
5: INTERSECTION DESIGN AND CONTROL

without left turn phase, it is apparent that the left turners have to depend on the available gaps in
the opposing through traffic. Given a sufficient length of left turn bay, the interaction between
left turners and through traffic can be avoided. If there is no left turn bay, the left-turning
vehicles are mixed with straight-ahead and right-naming vehicles (in ease of single lane) on the
same approach. This may have the following effects on the traffic flow:
a. Because they are delayed from turning left by other vehicles in the traffic stream, they
delay straight-ahead vehicles that may be following them.
b. The presence of left-turning vehicles in a particular lane tends to inhibit the use of this
lane by straight-ahead vehicles.
c. Those left-turning vehicles that remain in the intersection after the expiration of the green
period delay the start of the next phase until they have completed their left-turning
maneuver.

With fixed time left turn phase

As the volume of left turn traffic increases, providing additional phase for them seems
inevitable. The simplest Conn is a fixed time left turn phase. Fur any given traffic c Mimi with
the signals operating under fixed time control, the duration of the cycle affects the average delay
of vehicles passing through the intersection. Where the cycle time is very short, the proportion of
the cycle time occupied by the lost time in the intergreen period and by starting delays is high,
making the signal control inefficient and causing lengthy delays.
On the other hand, when the cycle lime is considerably long, waiting vehicles will clear
the stop line during the early pan of the green period and the only vehicles crossing during the
latter part of the green period will be those that subsequently arrive, often at longer headways. As
the discharge rate across the stop line is greatest when there is a queue on the approach, this also
results in inefficient operation.

With traffic-actuated left turn phase

For left turn traffic using traffic-actuated control. a special feature is that prior to the
beginning of green for left turn, the system checks whether there is a vehicle waiting, and if
found none, left turn phase is automatically skipped and the next phase begins. On the other
5: INTERSECTION DESIGN AND CONTROL

hand, if there is a vehicle waiting, then a minimum green period is given, and it may be extended
when there is continuous demand.
In the case of no exclusive left turn phase, there is usually enough time for left turners
during green if there is relatively light opposing traffic. However, as opposing traffic increases,
long queue may not be able to give the left turners the chance to clear the intersection through
available gaps. Or even if the opposing queue ends, the green time left with random flow may be
too short to clear all left turners waiting for gaps. Observation at the intersection will show that
those left turners leading the queue position themselves in the center of the intersection and turn
during the yellow period or the start of red interval. For high opposing traffic, this hangmen
period serves as the main source of left-turning traffic.
In the case of no exclusive left turn phase, it can be said that left turn capacity greatly
depends on the amount of traffic opposing the left turn movement. The left-turning vehicles must
cross the intersection through gaps or lags. A simple representation of this phenomenon is shown
in figure 5.26.
The condition on the opposing approach is shown in a, b, and c while the condition of the
left turn movement is shown in d. Briefly, the opposing traffic accumulates during red period;
then, at the beginning of green plus starting loss, the traffic starts to move. During normal
conditions, the queue can be cleared after some of the green time has been utilized and later the
flow on the opposing traffic returns to free flow condition wherein the vehicles arrive and depart
from the stop line in random fashion. From this point of time on, until the green time ends, the
waiting left turners assess the gaps available to them. Normally, the decision whether or not to
accept the available gap is made in the following manner:
a. No driver will accept a gap in the opposing stream that he or she belies will certainly
lead to a collision.
b. No driver gains admittance through intimidation of the opposing stream drivers.
5: INTERSECTION DESIGN AND CONTROL

In reality, assumption b is oftentimes violated. This may be due to the driver’s irritation
caused by the length of time he or she has been delayed in the queue.

5.7.13 Left Turn Lane Length

The usual approach in determining the required length of the storage lane or left turn bay
has been to assume a Poisson arrival distribution
and to compute the probability of the facility falling through overloading in not more than
specified percentages of cycles. If the storage to he provided is for k vehicles and the mean rate
of arrivals in an interval of time t is or, then

(5.13)
expresses the probability of the number of arrivals r in time t being greater than the storage
capacity k. A design based on this approach assumes that the storage provided allows through
traffic to proceed without interference and makes maximum use of the separate turning phase
through saturation flow away from the queue of vehicles in the storage lane.
5: INTERSECTION DESIGN AND CONTROL

This design approach which is frequently used by traffic engineers, generally forms the
basis for the storage length recommendation given by the American Association of State
Highway Officials (AASHO), which states. "At signalized intersections, the required storage
length depends on the cycle length, the signal phasing arrangement and the rate of arrivals and
departures of left turning vehicles. The storage length should be based on 1.5 to 2 times the
avenge number of vehicles that would store per cycle, predicated on the design volume."
Figure 5.27 shows a typical approach of an intersection when exclusive left turn lane is
provided, say, for k, vehicles. If the (k + 1)th vehicle arrives with intention of turning left, then it
has to occupy the adjacent lane, thereby blocking the through traffic in that particular lane.
resulting in a considerable reduction of the saturation flow rate of the through traffic. Also, the
same thing happens when there is heavy through traffic. If the through traffic queue length
exceeds that of the storage length, then these vehicles will block the use of the storage lane by
subsequent left-turning vehicles. And if this blocking of left-turning vehicles happens prior or
during the left turn phase for that approach, then only those vehicles stored in the left turn bay
can cross.

5.7.14 Effects of Left Turn Vehicles on Saturation Flow Rate

Given a sufficient length of left turn bay, interaction between left turners and through
traffic in the same approach is almost negligible. However, if the tail of right turn queue exceeds
that of the given storage length, bother usage of the through lane adjacent to the left turn lane is

Figure 5.27
Blockage of through traffic due to inadequate length of left turn bay
5: INTERSECTION DESIGN AND CONTROL

stopped. If the number of lanes for one approach is 2, for example, it is reduced to 1 and the
saturation flow is greatly reduced, too. When there is no exclusive left turn lane provided, the left
turners greatly affect the flow of the through traffic. Also, the effect becomes considerable the
more the number of left turners waiting at the approach. Once a through-bound driver sees left
turners in the innermost lane, he normally keeps away from that lane.
This reduction of saturation flow rate of through traffic can also be applied to the case
where there is exclusive left turn lane provided. However, the problem is a little different since
blocking occurs at the junction of the left turn lane and the adjacent through lane. If left turn
volume is high and the left turn lane is inadequate, it may be expected that the left turn queue
will pile up until the adjacent through lane is blocked. Normally, at the start of through green, the
leaders of the left turn queue proceed to the center of the intersection and wait for available gaps
or leave the intersection immediately at the start of the left turn phase. This somehow eases the
problem of interblocking at the junction of left turn and through lanes. After some time from the
start of through green, blocking may no longer occur as the left turners at the end of the queue
have proceeded to the left turn lane.

5.7.15 Coordination of Traffic Signals

TO provide progressive movement of traffic flow that has to pass through a series of
signalized intersections, some form of coordination has to be introduced to minimized delay.
Nowadays, a number of commercial software packages are available to achieve this goal. Some
of these package programs are very sophisticated and can handle large numbers of intersections
within the city center. Two well-known computer software are the SCATS and the SCOOT. The
SCATS is currently being used in Metro Manila and Metro Cebu.
While the said software are used for coordination of more elaborate subgroups, a manual
method can be used for simpler systems. Three methods that make use of simple time-distance
diagram will be presented here.

Simultaneous system
All signals display the same color indication this is commonly used when intersections
are closely spaced, such as those along Espana.
5: INTERSECTION DESIGN AND CONTROL

Figure 5.28
Simultaneous system

A disadvantage of this system is that drivers tend to increase speed in order to pass as
many intersections as possible.

Alternating system
At any given instant of time, the driver sees intersections ahead with alternating green
and red indications. This system normally works well when the intersections are far apart.

Figure 5.29
Alternating system
5: INTERSECTION DESIGN AND CONTROL

Progressive system
The starts of green are arranged in such a way that traffic flow is uninterrupted and
bandwidth is optimized. The method works well for one-way systems or when one direction of
flow is given more priority over the other.
For coordination, it is better to use one common cycle for all the intersections. However,
if distances are irregular, some intersections may use two times the common cycle. In practice,
this common cycle time is usually the cycle time of the key intersection (i.e., the intersection that
normally requires the longest cycle time).

Figure 5.30
Progressive system
5: INTERSECTION DESIGN AND CONTROL

Example 5.11
Consider the 4 intersections 400 m apart and with the given traffic data. Determine the
common cycle and the actual green times that will be used for coordination.

Each intersection will be controlled by two-phase signal in the absence of left-turning

vehicles. The cycle length required for each intersection is computed. The result is summarized
in the table below. Values of optimum cycle length are rounded off.
5: INTERSECTION DESIGN AND CONTROL

The key intersection is B, requiring the longest cycle (140 sec). For this intersection,
green times are allocated as follows:

For the other intersections, the maximum actual green for other intersections depends on
the minimum effective green for the side roads. This minimum effective green is computed as

(5.14)
The table shows how the required minimum green is given to the side roads so as to
maximum the green along the N-S direction.

Finally, the actual green and red (in seconds) that will be used for coordinating the N-S
direction are as follows:
5: INTERSECTION DESIGN AND CONTROL

5.7.16 Graphical Method of Coordinating Signalized Intersections

The Institute of Traffic Engineers’ traffic Engineering Handbook has a graphical method
of coordinating two-way systems when balance of traffic flow in both directions is desired. The
step-by-step procedure is discussed:
a. Preparation of the time-distance diagram. The locations of the intersections to be
coordinated are drawn to scale on the horizontal axis. A construction line is drawn
parallel to this axis.

b. Knowing beforehand the common cycle time and its components (effective green and
red periods), a green or red is centred on the construction line. (The figure shows a
red [bold line] being centred on this line.)
5: INTERSECTION DESIGN AND CONTROL

c. A line (line 1) with a slope corresponding to an assumed desired speed of progression

is drawn. This line represents the first vehicle in platoon that will cross the first
intersection.

d. Going to the next intersection, a red or a green is centered on the construction line,
with the start of green very close to the initial slopping line. In the figure a green is
centered on the line. The slopping line is then adjusted; this gives the new line (line
2).
5: INTERSECTION DESIGN AND CONTROL

e. The steps are repeated for the next intersections

f. Finally, a line is drawn parallel to line 3. This line represents the last vehicle in the
platoon that will be able to cross all intersections without stopping. Lines are also
drawn to represent the platoon going in the other direction. It can be seen that the
method procedures a balanced flow in both directions. The final desired speed can
also be estimated from the diagram (inverse of the slope of the line).

5.7.17 Offset
The difference in the start of green of two signalized intersections is called the offset.
Two types of definitions may be used:
5: INTERSECTION DESIGN AND CONTROL

a. Absolute offset
This would require one of the intersections to be defined as the key intersection.
Then, all offsets will be reckoned from it. O21 (read as “offset of 2 with respect to 1”)
and O31 are absolute offsets when intersection 1 is defined as the key intersection in
figure 5.31.
b. Relative offset
The offset of an intersection is defined with respect to an adjacent intersection O32
is a relative offset if intersection 1 is the key intersection.

Figure 5.31
Offset definition
5: INTERSECTION DESIGN AND CONTROL

PROBLEMS
1. Assuming that all movements are allowed in a five-leg intersection, determine the
number of vehicular crossing, merging, and diverging conflicts.
2. Given the T-intersection below:

(The numbers shown are traffic volume in pcu/hr and saturation flow rates in
pcu/hr of green time, respectively.)
a. Design appropriate phasing and draw critical movement diagram.
b. Determine optimum cycle.
c. Draw timing diagram.
d. Compute degree of congestion, X.
Use all-red = 2 sec.
3. Show that the degree of congestion of any critical movement is equal to the degree of
congestion of the intersection given by
5: INTERSECTION DESIGN AND CONTROL

4. For each approach of the intersection, allocate lanes for though and left and design
appropriate phasing so that the degree of congestion does not exceed 0.90.

5. Two intersections are 400 m apart. Common cycle time used is a 120 sec and green time
used for each intersection is 50 percent of the cycle. Determine if alternating or
simultaneous system is better if progression is desired for both directions. What will be
the recommended speed of progression?
6. For the given intersection, arrange the lanes (i.e., allocate lanes for through and left) of
each approach and design appropriate phasing so that the degree of congestion dies not
exceed 0.90.

Each approach can accommodate maximum of four lanes (through and left).
5: INTERSECTION DESIGN AND CONTROL

a. Calculate optimum cycle time.

b. Draw timing diagram.
c. Compute average delay per vehicle of approach N.
7. Given the traffic data for the one-way road shown below:

Assume the following for all intersections:

All-red: 2 sec Yellow: 3 sec
Starting loss: 2 sec Desired progression speed: 30 kph
Coordinate the 4 intersections. Draw time-distance diagram using appropriate
scale. Estimate bandwidth graphically.
5: INTERSECTION DESIGN AND CONTROL

8. A four-leg intersection has the traffic volume information as shown. Due to the large
number of pedestrians crossing in all directions (including diagonal movements), a
scramble phase (i.e., exclusive phase for pedestrians) of 20 sec after all phases are given
to vehicles has been planned.

a. Determine the following:

 cycle length
 green time allocation for each phase
b. Draw timing diagram.
9. A pedestrian crossing located at a mid-block has a cycle time c and is provided with
effective green g. Pedestrian arrival rare is q. Saturation flow of pedestrians is considered
to be very large in comparison with the arrival rate, i.e., almost all pedestrians cross at the
same time at the start of green. Show that the average pedestrian delay is

REFERENCES
 Akcelik, Rahmi. 1987. Traffic signal setting: Capacity and timing analysis. Australian Road
Research Board
 Allsop, Richard E. 1972. Delay at a fixed time traffic signal, I: Theoretical analysis.
Transportation Science 6, no.3.
 Drew, Donald R. 1968. Traffic flow theory and control. McGraw-Hill Inc.
5: INTERSECTION DESIGN AND CONTROL

 Evans, Henry K., 1950. Traffic engineering handbook. Second ed. New Haven, Connecticut:
Institute of Traffic Engineers.
 Federal Highway Administration. 1988. Manual on uniform traffic control devices for streets
and highways. US Department of Transportation.
 Japan Society of Traffic Engineers. 1988. The planning and design of at-grade intersections.
 Mannering, Fred L., and Walter P. Kilareski. 1990. Princicples of highway engineering and
traffic analysis. John Wiley and Sons.
 Organizations for Economic Cooperation and Development (OECD). 1978. Capacity of at-
grade junctions. Road Research Group, Paris.
 Salter, R. J. 1976. Highway traffic analysis and design. Revised edition. MacMillan Press.
 Sigua, R.G. 1984. A study on control of right turn traffic at signalized intersection. Master’s
thesis, Department of Civil Engineering, University of Tokyo.
 ------------- 1991. An expert system for intersection design and analysis. PhD diss.,
Department of Civil Engineering, University of Tokyo.
 Transportation Research Board. 2000. Highway capacity manual. National Research
Council.
 Webster, Frederick V. 1958. Traffic signal settings. Road Research Technical Paper No. 39.
 Wright, Paul H. 1958. Simulation of traffic at a 4-way stop intersection. Analysis of Traffic
Flow symposium. Michigan.