Transport Research Laboratory Overseas Development Administration

Old Wokingham Road 94 Victoria Street

Crowthorne, Berkshire, RG45 6AU London, SW1E 5JL

OVERSEAS ROAD NOTE 13

THE USE OF TRAFFIC SIGNALS IN DEVELOPING CITIES

Main Subject area: Urbanisation/Transport, energy efficiency

Theme: Urban Transport

Project title: The Use of Traffic Signals In Developing Cities

Project reference: ODA R6016

Copyright Transport Research Laboratory 1996. All rights reserved.

This document is an output from a project funded by the UK Overseas Development Administration (ODA) for the benefit of
developing countries. The views expressed are not necessarily those of the ODA.

Transport Research Foundation Group of Companies

Transport Research Foundation (a company limited by guarantee) trading as Transport Research Laboratory. Registered to England, Number 3011746. TRL
Limited Registered in England, Number 31.12272 Registered Offices Old Wokingham Road, Crowthorne, Berkshire, RG45 6AU.
ACKNOWLEDGEMENTS

This Note was prepared by A. Cannell of Transcraft Consultants, Curitiba, Brazil and G Gardner of the Overseas Centre,
Transport Research Laboratory (TRL). Useful advice and assistance was given by D. Singh and J. Cracknell.

First Published 1996

ISSN 0951-8987

OVERSEAS ROAD NOTES

Overseas Road Notes are prepared principally for road and transport authorities in countries receiving technical
assistance from the British Government. A limited number of copies is available to other organisations and to
individuals with an interest in roads overseas, and may be obtained from:

Transport Research Laboratory

Crowthorne, Berkshire, RG45 6AU
United Kingdom

Limited extracts from the text may be reproduced provided the source is acknowledged. For more extensive
reproduction, please write to:

Programme Director, Overseas

Centre, Transport Research
Laboratory.
 CONTENTS
Page

Page Time Settings 18

18
1. INTRODUCTION 1 Total Lost Time per Cycle (L)
18
2. CRITERIA AND WARRANTS Flow Factors
19
FOR SIGNAL INSTALLATION 2 Cycle Times
19
3. BASIC TRAFFIC COUNT SURVEYS 4 Green Times
20
4. JUNCTION DESIGN AND LAYOUT 6 Degree of Saturation
20
Typical Layouts 6 Junction Capacity Analysis
20
Siting of Signal Equipments 6 Traffic Signal Calculation Sheet
20
Approaches and Lanes 6 Stage/Phase Sequence Diagram
20
Signal Sequences 7 Check List for Signal Design
Signal Design Techniques 9 6. COORDINATION AND LINKING
OF TRAFFIC SIGNALS 23
Right Turning Vehicles 10
Early Cut Off 10 Simple Progressive System
Late Start 10 (Green Wave) 23
Pedestrian Facilities 10 Mechanisms for Linking Signals 25
No Pedestrian Signal 11 Cable-Linking 25
Full Pedestrian Stage 11 Cable-less Linking 25
Parallel Pedestrians 11 Fixed Time Coordinated Signals 25
Staggered Pedestrian Facility 11 Area Traffic Control (ATC) 25
Pedestrian Signal Displays 11 Fixed Time ATC Systems 26
Pedestrian Push Buttons. 12 Semi-Responsive Systems 26
Audible Warnings 12 Fully Responsive Control 26
Guard Rails 12 Equipment Testing 2
Pedestrian Signal Sequences and 7. THE TRANSYT PROGRAM 7
Timings 12 8. SYSTEM AND ECONOMIC ANALYSIS
OF TRAFFIC CONTROL 30
Vehicle-actuated (V.A.) Traffic Signals 12
9. SPECIFICATIONS 32
Stage demands 13
Traffic Signal Controllers 32
Stage extension 13
General Road Traffic Signals 32
Semi-vehicle-actuated signals 13
Inductive Loop Detectors 32
Traffic Signals on High Speed Roads 13
Associated Electrical Works 33
Speed-related Green Extensions 13
Visibility Requirements 13 Traffic Signal Controller Civil Works 33
Bus Priority 13
10. GLOSSARY 34
5. CALCULATION OF TRAFFIC SIGNAL
11. REFERENCES 36
TIMINGS - WEBSTER'S METHOD 16
Cycle 16
Intergreen Period 16
Minimum Green Period 16
Estimation of Saturation Flow 16
Width of Approach 16
Gradients 17
Traffic Composition 17
Turning Traffic 17
Parking, Waiting and Bus Stops 18
THE USE OF TRAFFIC SIGNALS IN DEVELOPING CITIES

1. INTRODUCTION
1.1 A traffic signal installation is a power-operated device
which informs motorists or pedestrians when they have the
right of way at a particular intersection.

1.2 The first traffic signal was installed in London in 1868

and used semaphore 'arms' together with red and green gas
lamps. Unfortunately, it exploded, putting an end to this sort of
control for 50 years.

1.3 However, in 1918 the first three coloured light signals

were installed in New York and in 1925 they started to be used
in Great Britain.

1.4 At the beginning of the 1930's an attempt at making the

signals more `intelligent', or vehicle responsive, was tried in
America, using microphones at the side of the road, requiring
drivers to sound their horns. This was obviously not too
popular and the first traffic detectors - electrical and pneumatic
- were invented.

1.5 Traffic signals are now used throughout the world, using
the three light signals of Green, red and amber. Also, by Figure 1 Conflict points at an intersection
convention, these are normally arranged vertically with the red
signal at the top and the green light at the bottom. This also
helps people who are colour blind - both drivers and pedestrians 1.10 There is no doubt that signals are one of the most
- to identify the differences between the lights. powerful tools for urban traffic control available to city
authorities and their correct installation can improve both
1.6 Traffic signals are used at intersections to reduce traffic flow and the safety of all road users. In comparison to
conflicts to a minimum by time sharing of right of way. This other traffic improvements, signals are also relatively low
actually reduces the capacity of the intersection, but greatly capital intensive and in recent years the advancement in
enhances safety. informatics and telecommunications has led to a new
generation of low cost controllers and systems that have made
1.7 Conflicts at intersections are illustrated in Figure 1 which modern signalling an even more attractive and powerful tool.
shows the potential conflict points at the junction of two roads,
both with two way traffic flows, at which all crossing and 1.11 Essentially, traffic signals form part of the "soft-
merging movements are permitted. ware" of a city as opposed to the roads and bridges that are
part of it's "hardware". As such they have the advantage of
1.8 With the provision of traffic signal control the number of being cheap and often the disadvantage of being so cheap
potential conflicts can be reduced from 64 to zero. that no local lobby is interested in them, especially when
city mayors fail to see the political advantages in changing
1.9 The object of this report is to give traffic engineers or an old signal for a new one.
technicians in the cities of the emerging world a brief
introduction to traffic signals, together with some practical 1.12 It is thus part of the traffic engineer's task to prove to
guidelines on how to use them to obtain good and safe results. city authorities that a modern and well designed traffic
signal system will bring real and visible benefits to the city.

1
2. CRITERIA AND WARRANTS FOR - where a schoolchildren crossing is present;
SIGNAL INSTALLATION - where there is a need to maintain progressive movement
of vehicles along an otherwise signalled route; and
2.1 When two or more traffic flows are competing for the - where there is a record of accidents of the type which
same road space at a junction, some form of control - or set of
rules - is needed to minimize delays and the risk of serious could be reduced by the use of traffic signals.
accidents. In some countries, a simple rule of preference states
that the traffic coming from the left (or right where there is
right-hand drive) has priority to enter the junction. As few 2.7 A rough and ready set of warrants might be:
people tend to know - or obey - this rule, unsignalled junctions
can come under "popular control" and users have to consider
that the larger vehicle, or the one that sounded the horn first, or traffic flows - when there is a minimum of 1000 pcu's per hour
a public transport bus, etc., may have priority. entering the junction during the peak hours.

2.2 This is obviously inefficient and dangerous, so with visibility - when drivers on the minor road have poor visibility
higher flows some form of stop or priority sign is used to for judging gaps.
inform to the user on one or more approaches that the other
road has right of way. At even higher flows this form of control accidents - when three or more accidents (collisions or
breaks down when the delay on the minor road becomes too pedestrians) are registered per year.
high, forming queues and forcing drivers to run the risk of
accepting gaps in the major road traffic that are too small for a 2.8 Figure 2, for example, shows the relationship between
safe crossing. At this point, time must be allocated for the right- major-road/minor-road flows and the type of control
of-way to traffic on the various approaches. recommended at a junction in the UK. For a major road flow of
20,000 pcu's per day and a minor road flow of 6,000, a
2.3 However, the introduction of traffic signals (or lights) roundabout would be a good solution for eliminating the
into a city often runs the risk of these equipments being conflicting traffic movements - if space were available. If,
considered a panacea for all traffic problems. The engineer or however, the junction is in a built-up area, then traffic signals
technician in charge of the traffic comes under political and probably represent the best solution.
popular pressure to install too many signals, thus leading to the
even greater risks of red-running - as the users `learn' to 2.9 It should be stressed, however, that traffic signals if
disrespect the red lights that they consider to be unnecessary. located or timed wrongly can INCREASE delays and accidents
and their maintenance and electrical supply represents an
2.4 To avoid this problem it is essential that the engineer or ongoing cost of around US$1000 to 2000 per year.
traffic department has a clear set of warrants to justify the use
of signals. 2.10 To minimize the need for signals, the road hierarchy
should try to conform to the network shown in figure 3, which
2.5 If possible, these warrants should be approved by the offers the most efficient and safe layout.
local government bodies (elected and executive) so that
requests for signals on sites that do not need them can be
refused according to pre-discussed rules - and not just on the
personalized decision of the head of the traffic depart-ment.

2.6 Traffic signals may be justified if, usually two, of the

following criteria are present:

- where there is a minimum major-street/minor-street

conflicting vehicle volume;

- where there may be need to interrupt continuous flow

on the major road to allow traffic to exit from the
minor road without excessive delay;

- where a minimum pedestrian volume conflicts with a

minimum vehicle volume;

Figure 2 UK practice for intersection control selection

based on combinations of traffic flow

2
2.11 A method of reducing conflicts on local distributors and 2.12 Figures 4 and 5 show how, in some cases, conflicting
access roads is to physically separate traffic flows, allowing flows may be avoided - provided that no economical or
access but avoiding the pressure to install new lights. environmental restrictions exist.

2.13 If, however, traffic lights are to be installed, the

engineer and police forces should be in agreement on how
the flows are controlled. In many developing cities, the
police will often take manual control, assuming that they
can reduce traffic queue lengths. Research has shown that
this is not true (Walker et al, 1988). Police are reluctant to
stop a traffic stream even when it is no longer saturated, as
shown in figure 6. It is preferable to allocate police to
control illegal parking, removal of breakdowns and enforce
driver behaviour.

Figure 3 Ideal urban road network

Figure 5 Eliminating the need for traffic

signals -"7 esquinas", Arequipa, Peru

Figure 6 Typical flow/saturation relationship for

Figure 4 Elimination of conflicts at a junction pair police control - inefficient use of the
end of the green period

3
3. BASIC TRAFFIC COUNT
SURVEYS

3.1 For each site where traffic signals are being contem-
plated it is fundamental to obtain adequate data on the traffic
flows at the junction. Normally, surveys would be carried out
during the peak hour periods. However, it may be important to
have a broad view of the flows in the city throughout a normal
working day, especially when Area Traffic Control or linked
signalling are being considered.

3.2 Traffic counts are likely to be divided into two types:

- all day counts (normally during 16 hours of a work day)

usually mid block on key roads, with the objective of
defining the duration of the peak periods and general
vehicle composition; and,

- specific junction counts carried out with the objective of

providing the data for evaluation and design of the
junctions.

3.3 The classification of vehicles might be cars, taxis, light

vans, trucks (heavy and medium) and public service vehicles.
In some cities it will be necessary to include motorcycles,
cycles or other common vehicle types. The counts should be
made in periods of about 15 minutes, during at least two
working days. If the counts are not similar (as demonstrated in
figure 8), then the counts should be repeated on another
working day.

A simple 16 hour survey form could look like figure 7.

3.4 Specific junction counts are aimed at providing the data

for detailed evaluation and design. The peak periods can be
identified from the all day (16h) counts and the junction counts
should be undertaken in the peaks - including the "shoulders"
just before and after the peaks. Unless a city is subject to
excessive congestion, this usually means a count period of
about two hours for each peak. If an ATC scheme is under
consideration, counts should also be carried out at weekends.

3.5 Each surveyor can usually manage to count two

Figure 7 Simplified traffic survey form
independent flows. For a simple junction involving two
one-way streets, two surveyors (normally temporary staff) will
be needed, as shown in figure 9. a survey is made during a widespread "gridlock"; reported by
the surveyors in terms of near zero flow on all approaches.
3.6 Each site should also be carefully checked to make sure
that pedestrian volumes during the peak hours that might 3.8 TRL ORN 11, "Traffic Surveys in Developing Cities"
require special phases are also considered. should be consulted for further reference.

3.7 Counts in congested areas often suffer from the spillback 3.9 The warrants used and/or approved by the city to
of upstream queues which means that surveyors will not count justify the installation of signals are likely to include
the real demand of the traffic that wants to go through the accidents. It must be stressed that an updated accident data
junction, but only the traffic that actually manages to pass. This base is essential for completing the traffic surveys.
can lead to the classic case, in which

4
 Figure 8 16 hour traffic count on Peru Street, Mendoza, Argentinia, During two working days

Figure 9 Survey forms for a simple junction of

two one-way streets
5
4. JUNCTION DESIGN AND right of way first. The secondary signal in this case should not
be placed beyond the nearside of the junction.
LAYOUT
Approaches and lanes
4.1 The aim of any junction layout is to provide for the safe
movement of traffic, both vehicular and pedestrian, without 4.7 It is essential that approaching drivers are made fully
undue delay or congestion. Various alternative layouts may be aware of the nature of the junction by adequate signing.
considered and the ultimate choice will be governed by such Carriageway markings and/or channelized islands should be
factors as the nature and volume of traffic using the junction, used to guide users on the correct path, and visibility should
the availability of land and the cost. not be impaired.

4.2 The overall capacity of a road network is limited by the 4.8 Approaches should be marked out in lanes. Lane widths
capacity of individual junctions. Failure to provide the correct at signalled junctions should normally be between 3 and 3.6m,
type of layout at one particular junction may result in accidents, although 2.7m is acceptable in some instances where speeds
congestion and delay to an extent which may impair the are low and there are few large vehicles (trucks or buses).
efficiency of the road system over a wide area.
4.9 On roads where land is available the saturation flow and
TYPICAL LAYOUTS capacity of an approach can be increased by widening the road
to the vicinity of the junction to provide more ahead lanes. An
4.3 The following descriptions of junction layout and design example of this is shown in figure 10. Another option, where
procedures are based mainly on UK practice. Other standards there are large turning movements is to divide the road space
are of course possible. For example, in the UK signals are available to favour the turning lanes, as shown in figure 11.
located on the kerb, at the roadside with the "primary" signal
close to the stop line. In many countries overhead signals on the 4.10 Perhaps the most important factor affecting the
"far side" of the junction are the norm. Both methods have their capacity of a junction approach is the need to avoid
merits, however, a country will generally have it's own obstruction to traffic flow, either temporary (a taxi or bus
standards and such standards have to be adopted in designs. stopping for passengers) or permanent (a parked car). Plate
The important requirement is that signals should be consistently 1. clearly shows the problem caused by a (very) long term
designed, located and operated throughout the city and clear parked car which has eliminated a lane of traffic. In a
unambiguous indications given to all road users. situation such as the example in plate 2, even the most
sophisticated traffic signals will not improve the traffic
flow.
SITING OF SIGNAL EQUIPMENTS

4.4 The minimum requirement is one traffic signal in-stalled

I m from the stopline, on the nearside of the carriage-way. If
possible a second primary signal should be installed if there is a
central island or divider, or more than three approach lanes.
Minimum visibility distances from the primary signals are
given in Table 1

TABLE 1: VISIBILITY DISTANCES

85 percentile visibility distance

approach speed (m)

50 km/h 70
60 km/h 95
70 km/h 125
85 km/h 165
100 km/h 225
Plate 1. Parked car obstructing the approach
4.5 A secondary signal is normally installed diagonally - a severe capacity loss
opposite the first primary signal, as shown in figure 12.

4.6 When the signal method of control contains a special

right turn phase, extreme care should be used in the siting of
secondary signals for the direction of flow which loses

6
 Figure 10 Flared junction approach Figure 11 Extra road space given to approaches

TABLE 2: APPROACH LANE WIDTHS

Approach Lane width (m)

width (m)
Lane 1 Lane 2 Lane 3 Lane 4

3-5 3.50
5.50 2.75 2.75
6.00 3.00 3.00
8.00 4.00 4.00
8.50 3.00 2.75 2.75
10.00 3.40 3.30 3.30
11.50 3.10 2.80 2.80 2.80

Lane 1 is nearest the kerb

SIGNAL SEQUENCES 4.12 Traffic control is by means of red, amber and green
signals, supplemented by additional green arrow light signals,
4.11 Each signal face normally has three vertical lights with a tram signals, etc.
nominal diameter of 200mrn. The height of the centre of the
green lens from the surface of the carriageway (where light 4.13 The signal sequence at junction traffic signals in British
signals are placed at the side of the carriageway) should be not practice countries is red, red + amber and green, amber and red.
less than 2.1 metres nor more than 3.5 metres. If signals are Most Panamerican standard countries, however, use the
placed over the carriageway, this distance should not be less sequence red, green, amber and red and some countries adopt
than 5.0 metres nor more than 9 metres. other variations, eg. flashing green in place of amber.

7
 Figure 12 Typical layout of a signalled controlled junction

8
 steps in which the junction control is varied. The other, phase
control, refers to the periods of time allocated to each traffic
stream.

4.22 In UK practice a phase is used to describe a set of traffic

movements which can take place simultaneously or the
sequence of signal indications received by such a set of
movements. A stage is that part of the cycle during which a
particular set of phases receives a green indication.

4.23 In USA based practice, a phase is that part of a cycle

allocated to any combination of traffic movements receiving
the right-of-way simultaneously during one or more intervals.
An interval is a period of time during which all signal
indications remain constant.

4.24 The cycle is the complete series of stages during which

all traffic movements are served in turn. The cycle time is the
sum of each of the stage times.
Plate 2. Street markets: a safety risk as well as
a huge capacity restraint
SIGNAL DESIGN TECHNIQUES

4.25 Conflicts are reduced at signal controlled junctions by

4.14 the red light signal indicates the prohibition that holding certain traffic streams stationary while others are
vehicular traffic shall not proceed beyond the stop line provided allowed to pass. To hold all streams and release each in turn
in conjunction with the light signals, or if the stop line is not would remove all conflicts but would not be satisfactory since
visible (or there is no stop line), beyond the light signals. delays to all traffic would be high and effective capacity of the
junction would be low.
4.15 the amber light signal when shown alone, indicates the
prohibition that vehicular traffic shall not proceed beyond the 4.26 The art of designing an installation is to reduce delay
stop line, or if the stop line is not visible (or there is no stop and increase capacity while still maintaining a high degree of
line), beyond the signals, except in the case of any vehicle safety.
which when the light signal first appears is so close to the stop
line or light signals that it cannot be safely stopped before 4.27 Reduction in total delay and improvement in capacity
passing the stop line or light signals. The time for the amber can be achieved by:
signal is normally fixed for the city or region at 3 or 4 seconds.
- utilizing the lowest practicable number of stages in
4.16 the red and amber light signals together indicate an any signal cycle.
imminent change from red to green. However the red light still
prohibits forward movement. - ensuring that each approach is capable of carrying the
maximum predicted traffic flow for that approach.
4.17 the green light signal means that traffic may proceed, if
safe to do so. - ensuring that the time allotted to each stage is appro-
priate to the actual traffic flow.
4.18 the green arrow signal indicates that traffic may
proceed only in the direction indicated by the arrow. - if appropriate, coordinating the control of adjacent
junctions to maintain the flow of traffic `platoons'.
4.19 a flashing amber signal in some countries means that
drivers must proceed with caution. Normally displayed on all - allowing simultaneous non-hooking right turns.
approaches with a frequency of 1 hertz (1 flash per second),
- separating left turn movements with an exit lane
this signal is sometimes used from midnight to 4 or 5 o'clock in
controlled only by a "give-way" priority sign.
towns with notorious night time red-running.
- where the degree of conflict is acceptable and move-
4.20 Pedestrian signals are red and green, either with a green ments can be executed safely with the exercise of due
walking man and a red standing man, or with "WALK/ DON'T care, a conflicting move may be accepted (e.g. a right
WALK" signs. turn on full green).

4.21 There are two alternative concepts used in describing the - restriction of movements, e.g. banned right turns,
control of traffic by means of light signals. One, known as where conflicting manoeuvres are forbidden.
stage control, is concerned with the sequential
- separation of traffic streams which conflict, assigning
them to different stages.

9
 - considering different stage sequences for different 4.33 The approach which is permitted to flow over two stages
times of the day. should have a three light primary signal. The secondary signal,
placed beyond the junction, should have four lights, including a
- providing extra lanes for turning traffic or flares on right turn arrow of 300mm diameter (in addition to the full
junction approaches. green signal) illuminated on the second stage when the
opposing traffic has been signalled to stop, as shown in figure
- combining the green periods for vehicles and pedes-
13 and 14.
trians when this can be done safely.
- providing two separate green periods in a cycle (re-
peated greens) for important movements.

4.28 As an example of these principles, figure 12. shows a

four arm junction with two stages with all movements
permitted. This is a very common junction and two stage
operation forms the basis of signalling techniques. Traffic on
opposite arms flows simultaneously, while traffic on the other
two arms is stopped. Each arm may have one or more lanes on
approach but the right turning traffic may impede vehicles
wishing to proceed over the junction if the road width is
restricted. Where there is a relatively minor right turn flow the Figure 13 Early cut off stage sequence
capacity of the junction is reduced by the road space occupied
by such traffic waiting to turn right and by the time which has
to be provided to this movement in the cycle. If the right turn
manoeuvre is removed then reduced delay and improved
capacity can be expected. An alternative route may often be
indicated to traffic before the junction is reached. Usually
motorists can turn left before the junction, make two right turns
to appear at the junction on the left hand arm (known as a `g'
turn). Alternatively motorists can pass through the junction,
turn left and make two further left turns to appear at the
junction on the left arm (known as a `q' turn). Such "q" and "g"
turns should be carefully evaluated as there will be increased
costs to set against savings injunction delay. In the case of "q"
turns, the use of the junction twice by former right turn traffic
may adversely affect junction capacity and thus delays and
operating costs.

RIGHT TURNING VEHICLES Figure 14 Green filter arrow for right turn
LATE START
4.29 The usual practice is for opposing right-turners to turn on
the nearside of each other. With this arrangement locking of 4.34 An alternative way of dealing with right turning traffic is
turning movement cannot occur but driver visibility may be to delay the start of the opposing traffic by a few seconds. This
restricted. method causes difficulty at the start of the following stage if
the right turn flow is heavy and the opposing traffic cannot
4.30 On high speed roads or where right turning movements establish precedence. For this reason a late start stage is usually
are heavy (above 300 pcu's/h), separately signalled and not recommended.
segregated lanes are strongly recommended.
4.35 When both right turn movements are heavy, another
4.31 Another very common situation is the four arm junction option available is to hold both right turns with a red signal
with three stages. The types of control are known as either early while the ahead and left turn traffic flows unhindered. All
cut-off or late start. traffic is then stopped before the right turn traffic on both
approaches is released together on the same stage. It is usual to
EARLY CUT OFF separate the right turn traffic onto exclusive lanes with separate
signals on each approach. This method should be employed on
4.32 To facilitate a heavy right turn movement from one high speed roads.
approach, the green time of the opposing approach can be cut
off some seconds before the approach with the right turn. PEDESTRIAN FACILITIES

4.36 When a traffic signal installation is being designed or

modified, the nature and extent of pedestrian flow has to

10
be taken into account as well as that of vehicular traffic. The
object of providing pedestrian facilities is to assist pedestrians
to cross in safety, with the minimum delay to both pedestrian
and vehicular traffic.

4.37 There area number of alternative methods of achieving

this aim and the engineer has to consider which of these
methods can be best applied to individual sites, knowing the
pedestrian flow pattern, degree of saturation and site layout.

4.38 Each junction should be considered on its own merits,

taking into account factors such as infirm or handicapped
pedestrians, junction capacity and any available accident
statistics.

4.39 If full pedestrian stages are new to the local traffic

culture, great care should be taken to introduce them only
when accident data and high pedestrian flows justify their
need.

NO PEDESTRIAN SIGNAL

4.40 The presence of traffic signals at an intersection provides

assistance to pedestrians in crossing the arms of a junction,
especially where refuges are available, and in many cases no
further facility is necessary. An extended all red period between
two traffic stages to assist pedestrians is not recommended. This
practice leads to increased delays to traffic and to
driverdisobedience since the extended period will always be
present even when there are no pedestrians.
Figure 15 Full pedestrian stage
FULL PEDESTRIAN STAGE

4.41 With this facility, all traffic is stopped while pedestrian

movement is signalled across all arms of the junction. This STAGGERED PEDESTRIAN FACILITY
method will cause delay to traffic. However, the stage can be
programmed only to operate during certain hours or by demand 4.44 Where carriageway widths permit, a large island in place
from push buttons. Where the crossing is across a dual of the normal refuge may be provided. Pedestrians can
carriageway, additional push buttons on the central reserve negotiate one half of the carriageway when traffic on that
should also be considered. approach is held on red at the junction signals. Normal
pedestrian signals are shown during this period. The other half
4.42 Although pedestrians may be allowed to cross any of the of the road is controlled by separate signals which are located
approaches to an intersection there will usually be one approach at the opposite end of the island. Normally the stagger should
upon which the pedestrian problem is most acute. The be at least one crossing width in order to alert pedestrians that
pedestrian stage should immediately follow the end of the the crossing is in two sections. A right-hand stagger may
vehicle stage on this approach. The signal sequence should be reduce junction intergreen times by placing approach stop lines
arranged to ensure that on termination of the pedestrian period, closer to a junction. A left-handed stagger, as shown in
the right of way will revert to a nominated stage in the absence figurel6, is normally preferred as pedestrians stepping on to the
of other demands. central refuge will turn towards the approaching traffic stream.

This is shown in figure 15. PEDESTRIAN SIGNAL DISPLAYS

PARALLEL PEDESTRIAN STAGES 4.45 Normally each signal face has two lights arranged
vertically (the upper red standing man and the lower green
4.43 Where it is possible to prohibit permanently some walking man) of 300mm nominal diameter. An alternative size
turning movements a combination of pedestrian and vehicle of 200mm nominal diameter may be used when specified.
stages can be installed. By virtue of banned turns, pedestrian
facilities can be provided across appropriate arms. In order to 4.46 The red stationary man, when illuminated by a steady
reduce the possibility of vehicles turning illegally, kerb radii light, indicates to a pedestrian that he should not cross or start to
should be kept as low as possible. cross the carriageway at the crossing.

11
 tone for the green walking man period and an intermittent tone
for the flashing green period.

4.51 Each proposal for use of audible signals at junctions

should be considered on individual merits and carefully
checked against real demand, safety aspects and potential risks,
technical feasibility of the equipment or supplier, local layout
and environment (these signals are not popular with nearby
residential blocks of flats).

4.52 An additional benefit to the visually handicapped can be

given by fixing metal plates with the street names in Braille
onto traffic signal posts in the vicinity of schools or other
buildings frequently used by them.

GUARD RAILS

4.53 It is desirable in some cases to restrict the crossing of

pedestrians to certain approaches at an intersection and guard
rails can be used to prevent pedestrians crossing at dangerous
places (for example where filtering traffic may be moving at
times unexpected by pedestrians). Guard rails should always be
provided on large islands where staggered pedestrian
movements are allowed. Normally minimum length of guard
rails provided at each side of a crossing should be 15m.

PEDESTRIAN SIGNAL SEQUENCE

AND TIMINGS
Figure 16 Left-handed stagger stage sequence
4.54 Pedestrian time should be sufficient to enable pe-
4.47 The green walking man signal, when illuminated by a destrians to cross the full width of the road with relative ease at
steady light, indicates to a pedestrian that he may cross the normal walking speed. An assumed walking speed of 1.2 m/s
camgeway at the crossing. for the measured crossing distance is satisfactory in
determining the minimum times. A staggered crossing can be
4.48 The green signal, when illuminated by an intermittent considered as two separate crossings.
light (flashing green man) indicates that a pedestrian who is
already on the crossing should proceed to complete the crossing 4.55 Normally, minimum green periods of less than 5
with reasonable speed; and/or a pedestrian who is not already seconds are considered too short and are not recommended.
on the crossing should not start to cross.
4.56 Provided that the above minimum requirements are met,
PEDESTRIAN PUSH BUTTONS the green period of a parallel pedestrian stage may be
determined by the predominant traffic flow running in parallel.
4.49 Pedestrian push buttons units mounted on signal posts
may be used for calling up pedestrian stages. Additional push 4.57 The vehicle clearing times before the start of all
buttons are also necessary on wide refuges where pedestrians pedestrian stages should be checked to ensure that the last
may be trapped at the end of the pedestrian stage. It is advisable vehicle clears the crossing by the time the pedestrian green
to have push buttons located at each side of the pedestrian signal is lit.
crossing, so that pedestrians approaching from either direction
can pass a push button before reaching the crossing. A summary of pedestrian facilities is given in table 3.

AUDIBLE WARNINGS VEHICLE-ACTUATED (V.A.)

TRAFFIC SIGNALS
4.50 Audible warnings, in the form of pulsed tones, are
intended for the benefit of visually handicapped pedestrians. 4.58 With vehicle -actuated (VA) signals the duration of
The set up consists of a post-mounted audible device which the green periods and the cycle time will vary in relation to the
emits different patterns of audible signal, representing different traffic flow into and through the controlled area. A vehicle-
pedestrian signal indications e.g. a slow hammering tone during actuated signal responds to demands recorded for
the red standing man period, a quicker

12
the various directions of flow. Once a green has been given to a TRAFFIC SIGNALS ON HIGH
particular direction of flow, the length of green for that
SPEED ROADS
direction will be extended until all the traffic has passed
through the junction, or the maximum green time for that
4.65 When traffic signals are installed on roads where the 85
direction has been reached.
percentile approach speed at a junction is between 60 km/h and
105 km/h on any arm, drivers have a difficult decision to make
4.59 Vehicle actuated signals will be most appropriate for
when green changes to amber: they are often faced with a
isolated junctions where coordination with other signals is not
choice between attempting to brake to a halt at the stop line, or
important and for locations with fluctuating light or medium
continuing at the same speed through the junction and clearing
traffic flows.
it safely.
STAGE DEMANDS 4.66 They may fail to achieve either, thus putting themselves
and others at great risk.
4.60 On the approach to a red signal, a green signal will be
demanded on the arrival of a vehicle on that approach. This 4.67 Because of the increased braking distances required at
demand is stored in the controller which will serve stages in high speeds, drivers need adequate warning that they are
cyclic order omitting any stages for which no demand has been approaching a signalled junction. High approach speeds also
received. Where it is essential that one stage must always result in drivers misjudging the lengths of gaps in opposing
follow another, the appearance of the first stage will traffic when making a right turn at the junction -again leading
automatically insert a demand for the second stage. to increased risk.
4.61 When a stage loses right of way on a maximum green 4.68 On high speed roads, the use of right turn clearance
period change, then a demand is inserted for a reversion to that phases should be avoided. Right turning movement, across
stage after other demands have been met. high speed flows should be channelized and controlled with a
separate vehicle phase, or preferably banned.
STAGE EXTENSION
SPEED-RELATED GREEN EXTENSIONS
4.62 When a green signal is displayed, the period for which it
is displayed may be extended by vehicles detected moving 4.69 To assist drivers and minimize risk it is necessary to
towards the signal. The purpose of this extension, or the sum of provide green extensions, the extensions being related to the 85
several extensions, is to permit vehicles to pass the stop line percentile approach speed. Normal approved vehicle detection
before the maximum green period is reached. equipment is used within 40m of the stop line on each approach
and in addition approved speed discrimination or speed
SEMI-VEHICLE-ACTUATED SIGNALS assessment equipment can be used.

4.63 With some semi-vehicle-actuated signals, detectors are 4.70 Advance warning signs are necessary on each
installed on the side roads only (i.e. not all approaches) and the approach, according to local or regional standards.
right-of-way normally rests with the main road, being
transferred immediately or at the end of a preset period to the 4.71 When the 85 percentile approach speed on any arm
side road when a vehicle passes over the side road detector. exceeds 105 km/h it is recommended that traffic signals
The green period on the side road can be extended in the should not be installed.
normal way by successive demands up to a preset maximum.
After right-of-way has been returned to the main road, it cannot
BUS PRIORITY
be taken away from the main road until the preset period has
expired.
4.72 The great majority of passengers in the cities of the
developing world travel by bus. Although these road users
4.64 Another modified form of V.A. signals is to operate one
normally have less political influence than the more affluent car
or more demand-dependent stages within a fixed cycle time.
owners, the traffic engineer should consider how to improve
The demand dependent stages which may consist of vehicle
bus flows at signalized junctions.
phases (such as right turn traffic, minor flows) or pedestrian
phases may be slapped or extended in accordance with the
4.73 The simplest form of priority is to guarantee that
prevailing situation detected. The advantage of this type of
saturation on the approaches most used by buses is kept as low
control is that a fixed cycle time can be maintained for linking
as possible, even if this means additional waiting times for the
with surrounding controllers.
other stages.

13
 Figure 17 Phase and stage sequence for early cut off operation

4.74 In ATC systems, the TRANSYT program (see section 7) - the use of pre-signals on the approaches to junctions.
permits bus flows to be treated separately thus providing These enable traffic queues to be relocated upstream
optimum settings for buses. of the junction and control traffic and bus flows to an
advance area so that all vehicles are able to clear the
4.75 rity to buses, not necessarily within ATC systems has been junction. (TRL ORN 12, 1993).
achieved at traffic signals by a number of methods. These
include:

- the selective detection of buses using on-bus trans-

ponders and detectors in the approaches to signals;

- the use of segregated lanes, exclusively for buses on

approaches to junctions, within which detectors are
installed to actuate the signals; and

14
TABLE 3: SUMMARY OF PEDESTRIAN FACILITIES

Type of facility Characteristics

No pedestrian signal - Traffic signals, even without signals for pedestrians, can help
pedestrians to cross by creating gaps in traffic streams.
- Especially applicable where there are refuges and
on one-way streets.

Full pedestrian stage - All traffic is stopped.

- Demanded from push buttons.
- More delay to vehicles than combined vehicle/pedestrian stages.

Parallel pedestrian stage - Combined vehicle/pedestrian stage often accompanied

by banned vehicle movements.
- Useful across one-way streets.

Staggered pedestrian facility - Pedestrians cross one half of the carriageway at a time.
- Large storage area in the centre of the carriageway required.
- Stagger preferably to face on-coming traffic.

Displaced pedestrian facility - For junctions close to capacity.

- The crossing point is situated away from the junction but within 50m.
- Normal staging arrangements as above apply.

15
5. CALCULATION OF TRAFFIC distance 'x' should be determined from the position of the
pedestrian crossing. Where pedestrians are losing right-of-way
SIGNAL TIMINGS - WEBSTER'S the start of the following stage should be delayed until the
METHOD crossing area is clear.

CYCLE MINIMUM GREEN PERIODS

5.1 A complete series of stages during which all traffic 5.8 Minimum Green Periods cannot be overridden by any
movements are served in turn is known as a cycle. The cycle demands, whether emanating from vehicles, manual control
time is the sum of each of the stage times. devices or received remotely from central computers or linked
controllers. Such a period is built into signal controllers. The
INTERGREEN PERIOD shortest minimum green period normally used for vehicle
stages is six to eight seconds.
5.2 The period between the end of the green display on
5.9 Site conditions may require a longer period where large
one phase and the start of the green display on the next
numbers of heavy vehicles have difficulty in starting, or the
phase gaining right-of-way is known as the intergreen
approach is on a steep gradient.
period. It comprises an amber display, red + amber display
and may also contain a period when the red signals are
5.10 Where pedestrians and traffic share the same stage,
shown to all approaches simultaneously. In some countries
minimum green times may be governed by the time required
thus intergreen period is composed solely of an amber signal
by pedestrians to clear the crossing.
and an all red period. With a five second intergreen the
amber and red + amber periods occur consecutively. Any
period over five seconds will include a period where red ESTIMATION OF SATURATION FLOW
signals are shown to all approaches simultaneously. (an all
red period). Safety requirements may dictate a longer period WIDTH OF APPROACH
to be given in the following circumstances
5.11 The Road Research Technical Paper No. 56 suggested
- to allow vehicles to clear the intersection when the that the Saturation Flow (S) be expressed in terms of passenger
distance across the junction is excessive. car units (pcu's) per hour and with no turning traffic or parked
vehicles;
- to improve safety on high speed roads.
S = 525w ,
- on roads where there are insufficient numbers of
right-turning traffic to justify provision of a separate where w is the width of the approach road in metres and
stage. 5.15<w <18.3m.

5.3 It should he noted that an intergreen period which is 5.12 For widths less than 5.15 m the following values may be
too short will be potentially dangerous but a period which is used:
too long is equally unsatisfactory since it may lead to delay,
frustration and lack of observation by drivers. A guide to w(m) 2.70 3.00 4.00 4.50 5.15
determining the length of the intergreen period is illustrated S(pcu/h) 1790 1850 1950 2250 2700
in Table 4.
5.13 Research and experience has shown that S may be more
5.4 A vehicle which passes over the stop line at the start accurately reflected by the number of lanes rather than the
of the amber display must be clear of the potential collision overall width of the approach - which tends to underestimate
point in relation to a vehicle starting at the onset of the flows in situations where narrow lanes of, say, 2.70m are used.
green of the following stage, when travelling at the normal
speed for the intersection. The distances AF and BF should 5.14 The correct value of S can be checked against the
be determined and those distances which give the highest formula:
difference used The recommended intergreen period can
then be determined. S = Si, (approx. 1790n - 100)

5.5 When East-West arms are losing right-of-way if AF- where n = max. no. of lanes (min 2.70m) and Si is the value of
BF is greater than CH-DH, then 'x' = AF-BF (or vice versa). S for each lane.

5.6 When North-South arms are losing right-of-way if 5.15 In practical terms this value is about 1790 pcu/h, except
DE-AE is greater than BG-CG, then x'= DE-AE (or vice for the nearside lane which is 100 units less. The higher value
versa). of 525w or 1790 - 100 should be used.

5.7 When the following stage is a pedestrian stage the

16
TABLE 4: CLEARANCE TIMES FOR INTERGREEN CALCULATION

Ahead Traffic

Distance 'x' (m) 9 10-18 19-27 28-36 37-46 47-54 55-64

Intergreen (seconds) 5 6 7 8 9 10 11

Turning Traffic

Distance 'x' (m) 9 10-13 14-20 21-27 28-34 35-40 41-45

Intergreen (seconds) 5 6 7 8 9 10 11

5.16 Provided that the junction is operating at under urban bus also assumes a vehicle of 10-12m. For motor cycles,
90%n of full capacity, the calculated green times or the lower value suggested would reflect a high proportion of
TRANSYT settings will still be valid even if local factors smaller types of vehicles, (under 125cc).
slightly alter the real values of S. However, for key
junctions that are operating at capacity there is no 5.20 It is worth remembering that although there has been a
alternative but to measure the saturation flow according lot of discussion on the changes of the value of S due to
to TRL Overseas Road Note 11. smaller, faster cars, etc., these factors tend to act equally on all
the approaches of a junction.
GRADIENTS
TURNING TRAFFIC
5.17 For each 1 % of uphill gradient the value of S should
be reduced by 3% up to a maximum of 30% reduction. For 5.21 When this traffic has an exclusive lane with no opposing
downhill gradients another 3 % should be added to the value flow, S may be given by the formulae:
of S (up to a max. of 15%) for every 1 % of gradient.
1800
TRAFFIC COMPOSITION-PASSENGER CAR UNITS S= (for one lane)
1+1.515/r
5.18 The effect of different vehicle types on the satura-tion
flow is eliminated by transforming all vehicles into standard 3000
car units, using a conversion factor, such as the values S= (for two lane)
suggested in Table 5. 1+1.515/r

5.19 The figures shown in Table 5 give an idea of the range where r = av. radius of turn
of values for different vehicle types in a number of countries.
The suggested values represent typical conditions in 5.22 In mixed traffic each left turning vehicle is equivalent to
developing cities, where heavy goods vehicles are normally 1.25 straight ahead vehicles, and each right turning vehicle is
larger than in Europe or Japan. The value for an equivalent to 1.75 straight ahead vehicles.

17
 TABLE 5: SOME TYPICAL PCU VALUES AT URBAN INTERSECTIONS

vehicle type England France Japan Indonesia India Cairo Chile Suggested
1966 1974 1974 1984 - 1985 1984 Values
Car 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
Mini-bus < 20pass. 1.0 - - 1.25 - 1.0 1.26 1.25
Motor Cycle 0.33 0.3 0.33 0.2 0.25 0.5 0.64 0.3
Heavy Goods 1.75 2.0 1.75 2.25 2.8 1.6 2.23 2.5
Bus>20pass. 2.25 2.0 1.75 2.62 3.6 2.5 1.52 2.5
Auto Rickshaw - - - 0.52 0.6 - - 0.5
Pedal Rickshaw - - - 0.93 1.4 - - 1.0
Pedal Cycle - 0.3 0.2 - 0.4 - - 0.3
Horse &Cart - - - - 2.6 4.0 - 3.0
Bullock Cart - - - - 11.2 4.0 - -

PARKING, WAITING AND BUS STOPS TIME SETTINGS

5 23 A parked car on an approach causes an effective loss in TOTAL LOST TIME PER CYCLE (L)
width, which can be expressed as:
5.27 With a 2 second red/amber start and 3 second amber
finish, the Intergreen Period for two conflicting stages is 5
loss = 1.68 - 0.9 (z - 7.62)m
seconds plus any all red time.
g
5.28 The lost time for a single phase during the green and
where z (>7.62) is the distance from the stop line of the parked
amber period is normally about 4 seconds, in other words, the
car and g the green time in seconds.
effective green time is thus the green time + 1. Then, per cycle:
5.24 In practice, the handling of parked or waiting vehicles in
Total lost time = 4*number of stages + all red
developing cities is extremely complex and involves a lot of
assumptions on the part of the engineer. Traffic signals are 5.29 Full pedestrian stages should be treated as all red tunes
generally put on the busiest junctions - and these are exactly the and included for the calculations.
favorite street corners for commerce and irregular bus stops.

5.25 Road signs and road markings may prohibit parking, but
FLOW FACTORS
enforcement is normally difficult or ineffective. Inmost cases
5.30 The flow factor `y' for each phase is given by the
the engineer will be left with the choice of reducing the
formula:
effective carriageway width by lm (for the odd parked car, a
bus stop or loading), by 2m (a row of parked cars with some
design flow for an approach (q)
double parking and occasional bus stop) or even 3m (next to a
y=
school where double parking is almost the rule) - or in
saturation flow for an approach (S)
situations such as the street market shown in Plate 2.

5.26 One of the other main causes of reduced saturation flow

5.31 Where more than one approach is operating during a
is the proximity of a signal controlled junction down-stream.
phase, the maximum `y' should be used. Where an early cut-off
This may result in the need for vehicles to slow down because
or late start is to be used in connection with a right turn the `y'
of poor linking or vehicles backing up from another congested
values for the right turn and for the approach with shortened
signalized junction downstream. The adverse effect of adjacent
time should be added together to represent one phase, unless the
signals on one another can be particularly serious during peak
straight on traffic on the same approach as the right turn has a
hours. In these circumstances site observations should be made
higher value, in which case this latter figure should be taken.
to evaluate the actual effective green usable by various vehicle
phases and to determine the necessary adjustments on signal
5.32 The sum of these higher `y' values for all the phases is
timings i.e. green splits and offsets.
the `Y' value, which is in fact a measure of the congestion, and
which will be used for calculating the optimum signal setting.

18
 Figure 18 Variation with time of discharge rate of a queue in a fully saturated green period

CYCLE TIMES 5.38 Even when pedestrian movement are low, a practical
upper limit to cycle times is around 120 seconds.
5.33 For an isolated signal installation, where the mean traffic
level is constant and where vehicle arrivals are at random, the GREEN TIMES
U.K. Transport Research Laboratory (TRL) has shown that the
optimum cycle time for minimum delay (Co) is given by: 5.39 Signal setting for the effective green periods (g) should
be in proportion to the y values on each approach, with an
1.5L + 5 allowance for lost time:
Co = secs.
1-Y g1 y1
= etc.
5.34 The cycle time which is just sufficient to pass the traffic g2 y2
(Cm) is given by:

L y (c – L)
Cm = secs. g=
1-Y Y

5.35 This is the minimum possible cycle time which may be Where:
associated with excessively long delays. In designing linked
signals a cycle time should be chosen which provides a margin g = effective green period
over this minimum cycle time for the key intersection. y = flow factor
G = actual green period
5.36 In practice it will be generally appropriate to choose a c = cycle time
practical cycle time (Cp) which will allow the installation to be
L = total lost time
loaded to 90 per cent of its capacity:
Y = sum of y flow factors
0.9L c-L = total effective green time
Cp = secs.
0.9 - Y Therefore:

5.37 Where pedestrian crossing volumes are high it will be y1 (c – L)

desirable to use as short a cycle time as practicable to minimize g1 =
delays imposed on pedestrians. In these cases, recommended Y
practice normally limits cycle times to below 90 seconds. Cycle
times lower than 50 seconds, on the other hand, tend to waste y2 (c – L)
too much of the cycle time in lost time. g2 =
Y

5.40 To simplify the calculations, for most cycle times of

around or above 60 seconds, the same value can be used for

19
both G and g. In other words, if the total lost time for the 5.46 Most of the useful formula have been incorporated so
junction is 8 seconds and the desired cycle time is 60 seconds, that the engineer may perform the operations directly, without
both G and g can be taken as 52 seconds. referring back to this guide. This particular model was
composed using a spreadsheet and can be adapted according to
5.41 If parallel pedestrian facilities are included in the local conditions and the program most suitable.
junction method of control, the minimum green times for the
minor movements could well be dictated by parallel pedestrian STAGE/PHASE SEQUENCE DIAGRAM
crossing green times. This could distort the green split
calculation and in situations where pedestrian signals are being 5.47 The method of signal control should be fully illustrated
introduced the engineer is faced with the dilemma of either by the Stage/Phase Sequence diagram, complete with the
reducing the minimum pedestrian time or oversaturating the following details:
junction capacity, in turn leading to disrespect of the pedestrian
signal. In many developing cities this problem still remains. - diagrammatic junction layout
- signals operation sequence
DEGREE OF SATURATION - design flows in pcu/h
- pcu. factor (if necessary), for converting unclassified
5.42 Degree of saturation (X) for individual approaches may counts from veh/h to pcu/h
also be expressed as; - intergreen periods required
- actual green times (G = g - 1 )
qc
X= 5.48 All traffic signal calculations should be regarded as a
gs good first estimate. There is no substitute for on-site checks
and the `fine tuning' that should take place once the
where: changes or installations have been implemented.

q = design flow 5.49 This section has outlined the "Webster" method of
c = cycle time traffic signal design, which, although manual, has great merit in
S = saturation flow that it is simple to apply and enables a clear understanding of
g = green time for approach the principles used. However, there have been great advances
in PC based programs for signal design, such as TRANSYT,
5.43 The degree of saturation should be the same for all the OSCADY, LINSIG, and others which enable calculations to be
predominant arms of an intersection when the signal timings are canned out rapidly and accurately.
at optimum settings and is given by the equation:
CHECK LIST FOR SIGNAL DESIGN
2Y
Xo =
Step 1 - Identify Traffic Flow Volumes
1+Y
Traffic flow volumes are identified, including turning
JUNCTION CAPACITY ANALYSIS movements.
5.44 The ultimate capacity of an intersection may be defined
as the maximum flow which can pass through the intersection
Step 2 - Identify Junction Layout, Lane
with the same relative flows on the various approaches and with Geometry and Site Characteristics
the existing proportions of turning traffic. Capacity will
normally increase as the cycle time increases, since the ratio of The junction layout, including lane geometry and site
lost to useful time decreases. In practice, for maximum reserve characteristics are identified.
capacity assessment, a maximum cycle time of 120 seconds
should be adopted. However it should be noted that for new It may be necessary, if revealed in Step 4 or Step 7, to modify
installations, the maximum operating cycle time should be the layout to cater for turning movements, pedestrians or to
limited to about 90 seconds. enhance capacity and/or safety.

TRAFFIC SIGNAL CALCULATION SHEET Step 3 - Identify Signal Phasing and Method of
Control
5.45 A `Traffic Signal Calculation Sheet' is shown as an aid
for performing traffic signal calculations.
The method of control to be used for analysis is identified.

20
21
Step 4 - Check Turning Movements and Step 11 - Determine Offset and Other Controller
Pedestrians Settings

Offset and other controller settings such as minimum green,

Adequate provision for turning movements and pedestrians
maximum green ,etc. are then finalized. Offsets for linking
should be checked. It may be identified at this stage that the
signals may be prepared with the aid of time-distance
assumed method of control would need adjustment before
diagrams. or programs such as TRANSYT. (see section 6).
continuing.

Adequate allowance in calculations for parallel pedestrian Step 12 - Prepare Documentation

minimum green crossing times should be made.
For record purposes, drawings showing the junction layout,
method of control, stage/phage diagram, traffic flow, etc. need
Step 5 - Estimate Saturation Flows to be prepared and maintained. Standard symbols should be
used wherever applicable.
The saturation flows for various approaches/movements are
identified. In critical cases the saturation flows for important Step 0 - The Proven Need for traffic signals at the
movements may have to be measured on site. junction according to approved warrants, should
always come first and be part of the documentation.
Step 6 - Compute Y, L,

The lost times, flow factors and sum of the critical flow factors
are computed.

Step 7 - Compute Reserve Capacity

The maximum reserve capacity of the intersection is then

calculated as a measure of operating performance. If this is not
satisfactory, then it may be necessary to go back to Step 2,
modify data and layouts and recalculate. A minimum provision
of 25% reserve capacity should be provided wherever possible
for new junctions. A lower standard may be adopted for
existing junctions where further improvement is restricted by
space limitations.

Step 8 - Compute Co, Cm, and Cp.

The optimum, minimum and practical cycle times for operating

the junction are then computed for further analysis, if
necessary.

Step 9 - Select C

It is then necessary to select a cycle time for operating the

intersection. Sometimes, for reasons of linking, the selected
cycle time may be different from the values calculated in the
previous step.

Step 10 - Compute Green Times, Degree of

Saturation

The green times of the various phases are then computed.

Degree of saturation may be computed as well if detailed
analysis of signal operation is required. If good linking to other
junctions requires a cycle time that results in very low degree
of saturation and a very high reserve capacity, consideration
should be given to double cycling this junction within the
linking group, i.e. running it at half the linking cycle time.

22
6. COORDINATION AND LINKING case, a desired speed should be used to ensure that the platoon
conforms to the legal limit.
OF TRAFFIC SIGNALS
6.5 The timings of the signals in a simple progressive system
6.1 Many cities in the developing world have grown to huge can be prepared with the aid of a tune-distance diagram,
dimensions in a very short period and lack the large scale road examples of which are shown in figures 19 and 20.
infrastructure needed for modern traffic, as well as the capital to
build it. This leads to the inefficient use of the existing road 6.6 On these diagrams, distances between junctions along
network, often in the form of extensive one-way systems, in the route are plotted along the abscissa (y axis) and the travel
turn leading to the need for a large number of traffic signals - times are plotted along the ordinate (x axis). The slope of
mostly fairly close to one another. diagonal lines represent the chosen speed of progression and
green stages of successive junctions are offset in time.
6.2 The effect of vehicles stopping at a signalised junction is Normally the problem is one of determining, by trial and error,
to form the vehicles into a queue behind the stop line. When the optimum through-band speed and width for a fixed cycle
this queue is released as the green is given, it will discharge time. For one-way roads, the green bands follow each other in
initially at its maximum rate (i.e. saturation flow) and move sequence. The driver, having passed one intersection, will then
forward as a `platoon'. If, as this platoon approaches another receive right of way at the others.
signal controlled intersection, its arrival is made to coincide
with the start of the green period, the vehicles will experience 6.7 When the flow of traffic is two-directional and where the
no delay at this new junction. If the platoon has to stop, a queue intersections are not equally spaced, the situation is more
may form leading to spillback, which in turn may block the complex and it may be necessary to come to a compromise on
upstream junction. progression between the two directions. It may also be
necessary to take into account other requirements such as
6.3 The objectives of coordination of signals are to: demands from cross-street traffic.

- prevent the queue of vehicles at one intersection from 6.8 The time-distance diagram method can be used to bias in
extending back and interfering with an upstream favour of a particular direction of flow e.g. to favour a heavy
junction. inbound flow in the morning peak at the expense of increased
delay to the fewer vehicles travelling in the opposite direction.
- increase the capacity of the linked route. The situation may be reversed for the evening peak.

- enhance driver comfort by offering less stops and 6.9 Cycle time for a coordinated signal system is normally
smoother flows in a controlled manner. dictated by the timings of a key junction, i.e. the junction which
is most heavily loaded Spare green time should be allocated as
- offer minimum overall delay for road users, reducing required to clear traffic turning into the main route from side
overall travel time. roads in order not to delay the through platoons.

- reduce fuel consumption - and hence pollution -within 6.10 To minimize congestion, opportunities for leaving the
the area. system should also be greater than for entering.

- impose on drivers a safer behaviour, as they normally 6.11 In heavy city centre traffic a design `speed' of about
bunch together at the speed designed for green l0m/s ( about 40km/h) usually gives good results. For suburban
coordination (or waves). Speeding is thus minimized, traffic, where traffic is lighter and signals are about 300m apart,
accident risk and severity are reduced as is the asso- a design velocity of 15m/s (about 60km/ h) can be used as a
ciated risk for pedestrians, as they can judge oncoming first estimate, provided this does not conflict with local speed
speeds with greater accuracy. restrictions.

6.12 On two-way roads, good coordination can usually be

SIMPLE PROGRESSIVE SYSTEM
obtained by using a common cycle time equivalent to twice the
(GREEN WAVE) average travel time between junctions.

6.4 The most commonly-used linking system works with a 6.13 For example, using a design speed of l0m/s on a two-way
cycle time common to all intersections and the signals are so road between nodes A.B.C, with links lengths AB+ 260m and
timed that the `go' periods are staggered in relation to each BC = 300m. The travel times are then 26 and 30 seconds with
other according to the road speed to give a 'progression' of an average of 28s. A good estimate of cycle time would then be
green periods along the road in both directions. Thus road speed 2x28 =56seconds, or in practice a cycle time of 55 to 65 would
should be considered "reasonable" by drivers; if speeding is be adequate (with the lower option possibly leading to higher
common before linking, then a measured speed will be too high speeds).
for safe operation. In this

23
 Figure 19 Co-ordinated signals for one-way traffic

Figure 20 Co-ordinated signals for two-way traffic

Red (amber omitted, only north-south phase shown

Green

24
MECHANISMS FOR LINKING SIGNALS 6.20 The main facilities provided by ATC Systems are

CABLE-LINKING - Optimized signal coordination. With the aid of

centralized computer control, signal settings can be
6.14 Local traffic signal controllers at intersections optimized on an area basis to provide minimum overall
working in a linked group may be cable-linked to a master delay and reduced journey times.
controller. The master controller generally ensures that the
local controllers, or `slave controllers' as they are termed, - Control flexibility. Changing traffic conditions can be
operate in synchronization by sending control pulses or catered for by vehicle actuation or predetermined multi-
instructions down the link cable to every slave controller. plan operation. The time settings of traffic signals can
also be altered very quickly by manual intervention at
the control centre, such as modifying the existing signal
CABLE-LESS LINKING
timing plan or replacing it with a new plan.
6.15 Linking of signals may also be achieved by cable-less
- Fault monitoring. One of the most important facilities
means such as radio, mains synchronization, etc. Originally
offered by ATC systems is the continuous monitoring
conceived to offer the benefits of wide area coordination while
of the operation of the traffic signal equipment linked to
avoiding the expensive costs of underground ducting, these
the computer. Any fault condition detected is reported
systems are now less competitive than the simpler ATC
to the Control Room immediately and fault repairs can
systems, as they do not allow for supervision of faults or such
be carried out quickly.
basic Responsive System facilities such as queue control.

FIXED-TIME COORDINATED SIGNALS - Priority for emergency and public transport vehi-
cles. For fire engines which always start from a certain
6.16 These systems are based on assumptions that traffic fire station, special plans may be prepared for
flows are repetitive over a weekly (24h, 7day) cycle and that predetermined `preferred routes' and stored in the
the appropriate predetermined signal settings can be prepared central computer. Priority arrangements can be given to
to cope with predictable traffic flows. The traffic plans, or public transport routes, busways, etc.
signals settings can be selected according to the day of week
and time of day. These systems require controllers with - Accident reduction. ATC systems improve road safety,
synchronized internal clocks and, typically, a minimum of 6 especially in the traffic conditions of some developing
plans to choose from. cities. Table 6 shows some results of traffic accidents
before and after the construction of an ATC system.
6.17 These would normally include plans for the:

- Morning Peak. TABLE 6: ACCIDENTS IN ATC SYSTEMS

-
- Off Peak (work day). ATC System Accident
- Reduction (%)
- Evening Peak.
- Toronto 13
- Off Peak (night, holiday or Sunday). Atlanta 24
- Witchita Falls 8.5
- Nighttime Yellow Flashing, etc. Sydney 20
Glasgow 14
AREA TRAFFIC CONTROL (ATC) Paramatta 8
West London 18
6.18 Area Traffic Control, (ATC) is the centralized control of Rochester 20
traffic signals on an area-wide basis using micro-processor and Curitiba 24
computer technology.

6.19 Usually the traffic controllers on street are linked to one 6.21 On certain corridors these improvements may be even
or more central computers in the control centre, via data greater. For a well designed system on downtown streets
transmission cables. The cable network can either be provided together with suburban corridors, an accident reduction of
as a dedicated network or private circuits leased from the about 20% (all types) may be expected.
telephone company (or a mixture of both, according to cost).
Urban Traffic Control, (UTC) involves central coordination of
signals as an ATC, but will include other facilities such as car
park space control and variable message signing.

25
 FIXED TIME ATC SYSTEMS examining all possible alternatives, during the evaluation of
any ATC or UTC system.
6.22 These ATC systems operate on a strategy of fixed time
traffic signal plans for each controller. Signal timings and plans SEMI-RESPONSIVE SYSTEMS
for junctions alter in accordance with a preset timetable held in
the central computer or in a microprocessor inside the 6.27 To overcome some of the rigidities inherent in the fully
controller itself. These systems work very well for traffic fixed time strategy, whilst avoiding some of the sophistication
patterns which are predictable and which change quite slowly. and complexities of fully responsive strategies, some degree of
Traffic signal timings are developed from historical traffic flow traffic responsive control can be introduced into the fixed time
data collected for the junctions and the timings require updating philosophy.
periodically, depending on the traffic fluctuations at each site.
6.28 This can take two main forms
6.23 For large urban areas with possibly several hundred
traffic signal controlled intersections, a large system, complete - Using strategic detectors to detect fluctuations in traffic
with a special computer, monitor rooms and possibly closed flow, the system automatically changes from one fixed
circuit television is an entirely appropriate arrangement. time traffic plan to another. Basically, this is equivalent
Usually this type of maxi system will demand (and can justify) to the introduction of a flexible timetable into the
its own dedicated team of hardware and systems engineers, system.
traffic engineers and technicians and system operators.
- Installing vehicle presence detectors and pedestrian push
6.24 For medium-sized cities in the developing world, buttons at junctions and introducing a VA 'window' into
however, the benefits accruing from the installation of a maxi the traffic signal timing plans. The difficulty with fully
system cannot be justified in terms of the initial costs and staff traffic actuated signals is that the concept of signal
resources demanded to run it. Experience has revealed that linking within a fixed cycle time breaks down under
fixed time systems are cheap and simple to install and maintain fully traffic-actuated control. The VA-window concept
and recent developments in electronic technology have made allows a certain amount of 'local discretion' in the traffic
the Mini ATC, or Compact ATC systems almost as efficient as signal plan. This permits the controllers at certain times
the traditional 'centralized' systems. The features of these in the signal cycle to decide whether to introduce a
systems are fairly standard, and can be summarized as: certain stage or whether to remain on the stage running,
or to control the extension of queues by allowing for
- Computer and data transmission system in a stand- green extensions, etc.
alone unit which does not require a specially controlled
environment. 6.29 The former is very effective in overcoming difficulties
in defining accurately plan change points. Also, as only a
- "off the shelf" system operating software. relatively small number of detectors may be required, a heavy
ongoing maintenance commitment on detectors is unnecessary.
- automatic printout of faults; can be left to run virtually And in developing cities detectors are often a problem.
unattended.
6.30 The second type of system can be very effective in
- usually these systems are purchased as a complete coping with minor fluctuations in traffic flow, queues and with
package to a standard specification. This minimizes the the problems of light vehicle and pedestrian flows in the off
staff input at the outset of the scheme. peak period. Though semi-responsive systems do require more
complex signal controllers on street.
6.25 Standardization usually means that the costs in-volved to
the purchasing of such a system are modest. As the system FULLY RESPONSIVE CONTROL
requires no special architectural/environmental arrangements,
and will virtually 'run itself,' the staff input on the part of the 6.31 The logical extension of computerised traffic control
purchaser/local administration is also small. Much of the philosophy is the fully automated system, where the computer
flexibility and monitoring facilities of the maxi systems are, automatically calculates and adjusts signal timings to suit
however, maintained and signal timings can be kept up to date actual traffic conditions on street.
by a small team, possibly on a part-time basis in addition to
normal traffic engineering duties. 6.32 The SCAT (Sydney Coordinated Adaptive Traffic)
System is based on a limited distributive intelligence system
6.26 One of the major difficulties and costs of ATC with a main computer performing certain monitoring functions
systems is in installing an adequate data transmission and several satellite computers controlling a group of
network. This item should be given special attention, controllers on street.

6.33 All intersections within a subsystem operate on a

common cycle length and are equipped with inductive loop
vehicle detectors on all approaches. These are located in

26
each lane immediately in advance of the stopline and perform 6.41 The curve shown in figure 21 represents the cost-benefit
the dual functions of providing traffic flow data for strategic relationship for the various options of control out-lined in this
control and local or `tactical' vehicle actuation. section. It should be noted that the principal gains are obtained
with less complex measures and that high-tech solutions -
6.34 The SCOOT (Split Cycle Offset Optimization although extremely useful options in big cities - are normally
Technique) System consists a number of SCOOT cells or only recommendable as a final stage in the ongoing process of
computers, each cell being able to control up to 60 junctions, traffic signal improvements.
handling input data from up to 256 vehicle counting detectors
on street. Unlike the SCAT system, the SCOOT detectors are
placed as far upstream from the approach to the junction as
possible and are then calibrated to strike a balance between
flow and occupancy.

6.35 In normal operation SCOOT estimates whether any

advantage is to be gained by altering the timings. If an
advantage is predicted then one or more of the timings are
changed by small amounts. By this means frequent, but small,
changes allow the signal timings to match fluctuations in traffic
demand. There are no large and abrupt changes in signal
timings, although overtime, major changes in splits, cycle time
and offset can occur.

6.36 It must be remembered that in order to operate any

fully responsive system at least 80% of detectors need to be
fully operational. This requires constant maintenance and
adequate continuous funding --exactly the areas which are
often a problem for transport projects in the developing
world.

DUET – DIAL-UP EQUIPMENT TESTING

6.37 It is arguable whether the greater benefit obtained from

an ATC system is derived from signal coordination or from the Figure 21 Cost/benefit relationship for different
ability to monitor continuously the performance of signal types of signal co-ordination
equipment. For signals in remote locations or where the cost of
a dedicated ATC telecommunication network is not justified,
DUET may provide a compromise solution. With the aid of this
technology the equipment can be interrogated over public
telephone lines as and when necessary.

6.38 Each local controller is connected to the public telephone

system. Communication is achieved by dialing a telephone
number from the Central Office, which connects the local
controller via modem to equipment in the Central Office for
testing/fault reporting purposes.

6.39 This arrangement avoids the need for a dedicated private

circuit but provides only a limited fault monitoring facility as
equipment faults will only be detected when the interrogation
takes place, which may be some time after the fault has
occurred.

6.40 An alternative is for the local controller to have the

ability to `ring-up' the Central Office when a fault is detected.
At the same time, it should be possible to change and load
plans from the Central Control Room via a DUET setup.

27
 7. THE TRANSYT PROGRAM as the common cycle time for the network and the cost of
delays to vehicles. Signal optimisation data allows the user to
specify what type of optimisation is to be carried out. Link
7.1 The TRANSYT program was developed by the U.K. data requires the user to specify information such as the flow
Transport and Road Research Laboratory in a series of and saturation flow on a link, as well as during which stages
experimental applications in Glasgow, Scotland, where an ATC the link receives a green signal. The node data requires the
System had been installed. user to specify the staging of the signals at each node,
minimum greens and intergreens.
7.2 The network being modelled in TRANSYT is repre-
sented by `nodes' inter-connected with `links'. Each signalised 7.6 This data is resumed in the input table as follows:
junction is represented by a node; each distinct one-way traffic
stream leading to a node is represented by a link. TRANSYT
takes the flows on each link and models traffic behaviour from
this.

7.3 Figure 22 represents a small network of three junctions

together with the traffic flows for one of the peak hours.

7.7 TRANSYT will produce an optimised set of signal

timings for each node within a network. The signal timings
are calculated by a method which works by varying both the
signal offsets and the duration of the individual stage green
times. Together, these signal timings result in delay being
reduced to a minimum within the network.

7.8 The program can also be used to advise on the best

cycle time for a network. This area of TRANSYT will also
advise the user about the benefits to be gained from double-
cycling any of the nodes.

7.9 For the optimised signal timings calculated by

TRANSYT, a table of results, relating to performance and
delay on each link, is given. From this table the user can
deduce how saturated each link is, the potential size of any
Figure 22 Simple road network and traffic flows queues, the cruise time for vehicles, the amount of delay each
vehicle can expect to suffer, as well as a series of cost related
7.4 This is then transformed into a "TRANSYT Dia- indices in terms of fuel consumption and delays.
gram" such as figure 23; note that turning movements are given
separate link numbers. 7.10 This program has been immensely successful and is
recognized as one of the most important traffic signal tool
available to engineers. Versions are now available for use in
P.C.'s and these also allow increased emphasis to be attached
to key links in the network and the whole process can, if
desired, be biased either towards minimum delay or
minimum stops to vehicles. The average speed within the
network and total fuel consumption are also given and these
provide the basic data for evaluating the operating benefits to
be gained from a new plan or system. Special routines are
also available to model bus movements separately and aid in
the selection of optimum cycletimes.

7.11 In practice, the TRANSYT Performance Index

may be improved in "Grid" type networks by building
into the initial offsets simple green waves on the main
streets of the network. The output is then normally a
much improved version of the desired coordination. If left
Figure 23 Transyt diagram of the network to optimize from an all zero setup, the program may opt,
when signals are close together, for a solution of the `open
7.5 TRANSYT requires the user to input a variety of all lights north-south, then east-west'. This leads to
information about the network. The data required is basically of speeding as drivers note that all the lights along the road
four types; network data, signal optimisation data, link data, go green at the same time.
and node data. Network data includes such things

28
TABLE 7: TRANSYT OUTPUT FOR A SIMPLE NETWORK

1 LINK FLOW SAT DEGREE MEAR TIMES --------------DELAY--------------- --------STOPS-------- -------QUEUE-------- PERFORMANCE EXIT GREEN TINES
NUMBER INTO FLOW OF PER PCU UNIFORM RANDOM+ COST MEAN COST MEAN INDEX. NODE START START
LINK SAT CRUISE OVERSAT OF STOPS OF MAX. AVERAGE WEIGHTED SUM END ENE
DELAY (U+R+O.HEAN Q) DELAY /PCU STOPS EXCESS OF ( ) VALUES 1ST 2ND
(PCU/H) (PCU/H) (%) (SEC) (SEC) (PCU-H/H) ($/H) (%) ($/H) (PCU) (PCU) ($/H) (SECONDS)

10 1200 2400 87 10 25 5.3 + 3.1 ( 68.1) 86 ( 9.9) 27 + 77.9 1 38 89

11 550 35005 81 8 31 3.6 + 1.2 ( 38.7) 98 ( 5.2) 24 43.8 1 4 33
12 400 11L 81 8 24 1.7 + 0.9 ( 21.3) 94 ( 3.6) 24 24.9 1 4 33
20 1150 2600 87 10 29 6.2 + 3.1 ( 74.6) 91 ( 10.0) 28 + 84.5 2 42 87
21 1101 38005 86 12 31 7.2 + 2.4 ( 77.3) 87 ( 9.1) 31 86.4 2 2 37
22 199 21L 86 12 39 1.7 + 0.4 ( 17.4) 108 ( 2.1) 31 19.4 2 2 37
30 500 1600 76 10 34 3.2 + 1.6 ( 37.9) 92 ( 4.4) 12 42.2 3 12 48
31 1000 36005 67 13 11 2.2 + 0.8 ( 24.1) 38 ( 3.7) 13 27.8 3 53 7
32 200 31L 67 13 5 0.1 + 0.2 ( 2.4) 22 ( 0.4) 13 2.8 3 53 7
39 100 1600 15 10 20 0.5 + 0.1 ( 4.4) 61 ( 0.6) 2 5.0 3 12 48
40 800 1600 82 10 23 3.0 + 2.2 ( 42.0) 80 ( 6.1) 18 + 48.1 4 63 27
41 550 3200S 73 10 27 3.1 + 1.0 ; 33.1) 90 4.73 17 37.8 4 32 58
42 150 41L 73 10 16 0.4 + 0.3 ( 5.3) 81 ( 1.2) 17 6.5 4 32 58
49 100 1600 10 10 9 0.2 + 0.1 ( 2.1) 40 ( 0.4) 1 2.5 4 63 27
TOTAL TOTAL MEAN TOTAL TOTAL TOTAL TOTAL PENALTY TOTAL
DISTANCE TIME JOURNEY UNIFORM RANDOM+ COST COST FOR PERFORMANCE
TRAVELLED SPENT SPEED DELAY OVERSAT OF OF EXCESS INDEX
DELAY DELAY STOPS QUEUES
(PCU-KH/H) (PCU-H/H) (EM/H) (PCU-H/H)(PCU-H/H) ($/H) (S/H) ($/H) (S/H)
843.0 79.8 10.6 38.2 17.5 ( 448.6) + ( 61.2) + ( 0.0) $09.8 TOTALS
******************************************************************************************************************
CRUISE DELAY STOPS TOTALS
LITRES PER HOUR LITRES PER HOUR LITRES PER HOUR LITRESPER HOUR
FUEL CONSUMPTION PREDICTIONS 59.7 + 78.0 + 60.7 = 198.3

NO. OF ENTRIES TO SUBPT = 10

NO. OF LINKS RECALCULATED= 115

29
 8. SYSTEM AND ECONOMIC 8.4 An analysis of existing traffic signals, with or without
coordination, plus those sites where counts or other data
ANALYSIS OF TRAFFIC suggest the installation of signals, using the above criteria will
CONTROL tend to show the areas or sub-areas that would probably benefit
most from a modern traffic signal scheme, such as those
indicated in figure 34.
8.1 In many cities throughout the developing world there is
enormous scope for introducing modern traffic controllers 8.5 Obviously, if the city has a population of several mullion
capable of offering linked signal timings compatible with the with a large and fast growing fleet of cars and buses, then some
different traffic volumes found during the day. The TRANSYT form of responsive control, such as SCOOT should be
Program can now be run on a 486 PC and similar advances in considered. In the city centre, important isolated junctions can
micro processors have given local controllers the "intelligence" also benefit from the application of an intelligent vehicle
that used to occupy large main frame computers, thus responsive system, such as MOVA (TRL. 1993.). Most cities in
simplifying the need for expensive telecommunication the developing world, however, will obtain the highest benefits
networks. from updating their traffic control system for a simpler model
using prefixed timings.
8.2 ATC systems improve road safety, they no longer
require expensive "command centres", with a large public 8.6 In this case, TRANSYT is a good tool for evaluating a
service staff and they normally have exceptionally high proposal.
Benefit/Cost ratios that make them potentially attractive to
international lending agencies and banks. 8.7 Having arrived at the study area (or areas) a TRANSYT
simulation for each area should be carried out, using real traffic
8.3 Warrants for an ATC System vary from city to city and data for each junction. If this data is not available, then new
there are no hard and fast rules that can be followed, however, counts have to be made.
some basic guidelines would cover:
8.8 If the number of junctions in the initial proposal leads to
- the city centre, where there is a high density of signals a cost that is unacceptable to the city authorities, then the
difficult task of reducing the proposal has to be undertaken,
- the streets used by public transport vehicles eliminating (or transferring to a later stage) those junctions or
sub-areas that show least benefits - providing that neither the
- one-way suburban roads signals are up to 600m apart safety nor the basic concept of the system is impaired.

- the main corridors and on the side streets up to 300m 8.9 TRANSYT simulations of these new areas, scaled to
from the corridor meet budget requirements, will then provide

- the access roads to the city from major highways - the existing operating costs and travel delays (using
existing flows and timings) for each peak period
- isolated junctions (using DUET or similar) that are
overloaded during peak hours.

City centre

Figure 24 Definition of ATC system sub-areas

30
 - optimized operating costs and delays (using the 8.15 DUET systems require the cooperation of the local
optimizing process), for each peak period telephone service - normally by contracting the lines, including
their maintenance and installation costs. This requires a written
- existing operating costs and delays for the off peaks proposal providing annual costs per junction.

- optimized operating costs and delays for the off 8.16 Software, engineering and system maintenance costs
peaks. vary greatly from city to city according to the size and
specifications of the proposed system. However, an additional
8.10 This same process can be repeated for a three, five and 30-40%o of the total would give a rough estimate.
ten year horizon, using the relevant traffic growth factors
(based on recent traffic or vehicle growth data). The yearly time 8.17 Considering the speed of advances in electronics and
and operating benefits of the proposed system can then be telecommunications a useful life of 10 years can be expected
estimated over the project's useful life. for ATC/UTC projects, allowing for a residual value of 20% of
the total project cost after year 10.
8.11 According to Table 6, a 20% reduction on the accident
costs within the proposed area(s) could also be included. 8.18 For two projects in Latin America, both nonresponsive
using prefixed timings, the economic data for the systems can
8.12 In developing cities it should be remembered that bus be summarized as:
(public transport) normally plays a major role and so these
travel benefits should be calculated, if possible, separately. The 8.19 Project a), Cost = US$ 5,794,000. Benefit/Cost Ratio =
value of time savings is always a delicate point - especially as 2.41. Internal Rate of Return = 39.88
these are aggregated in small amounts for a very large number
of people. However, it is standard practice to allocate a 8.20 Project b) Cost = US$ 5,000,000. Annual time savings =
reasonable value to these for economic evaluation. US$ 2.000.000. Annual fuel savings (excluding national and
local taxes) = US$ 3.709.548. Annual accident cost savings =
8.13 On major public transport trunk routes, the extra fleet US$ 2.640.000
capacity gained during the peak hours may also be relevant and
could be included as a real benefit if this means the 8.21 In both cases it is clear that projects of this kind have an
postponement of additional fleet capacity during the project's extremely high economic return. In fact, when a city proposes
lifespan. to upgrade its traffic signal system from isolated junctions with
just one cycle time and plan to a simple ATC network it is no
8.14 Table 8 below shows the costs of installing a new exaggeration to claim that the scheme will pay for itself in a
traffic signal layout to different parts of the world. A rough question of months.
estimate of junction cost for a modern, simple ATC System
would be around US$20.000 - excluding the costs of data
transmission and any detectors.

TABLE 8: ESTIMATES OF TRAFFIC SIGNAL INSTALLATION COSTS PER JUNCTION

city pop. junctions cost/simple cost/complex

with junction junction
(106) signals (US$) (US$)

São Paulo 12.0 4000 4.150 36.000

Santiago 4.5 980 14.300 28.600
Curitiba 1.6 350 12.000 18.000
Nairobi 1.5 27 18.000 26.500
Harare 1.0 152 8.000 16.000
Bombay 11.0 305 12.000 18.000
Istanbul 4.0 300 14.000 55.000
Islamabad 0.4 41 10.450 13.550

31
 9. SPECIFICATIONS - capability of incorporating pedestrian push buttons,
pedestrian wait indicators and pedestrian audible signals.
9.1 Most countries have developed their own national
specifications for traffic signal materials, based either on local 9.10 Most controllers offer a minimum setup of 4 stages
experience and production or on international standards used in and/or 8 phases, allowing for expansion to 8 stages and/or 16
the larger markets of North America, Japan, European phases, without the need to change the cabinet or internal
Community, etc. Detailed specifications of equipment and architecture of the equipment.
services are beyond the scope of thus report, however, some
basic requirements can be resumed as follows GENERAL ROAD TRAFFIC SIGNALS

TRAFFIC SIGNAL CONTROLLERS 9.11 Vehicular signals should contain three optical systems
arranged vertically each having a nominal diameter of 200mm.
9.2 The increasing demand for better traffic control
mechanisms and the advancements in technology have 9.12 Pedestrian signals should contain two optical systems
introduced the microprocessor-based traffic signal controller - arranged vertically and shall normally each have a nominal
now a fairly standard, cheap and reliable form of control. diameter of 300mm - an alternative size of 200imm nominal
diameter may also be used where necessary.
9.3 A major difference arising from the use of
microprocessors is that the operating parameters such as 9.13 Where an optical system incorporating a green arrow is
timings are stored in memory rather than by switches or used, it should normally have a nominal diameter of 300nim -
potentiometers. Changes in operation are accomplished by the an alternative size of 200mm nominal diameter may also be
use of hand-held terminals, although safety requirements are used if necessary.
met by storing the basic site configuration data and control
program on programmable read only memories (PROMS), 9.14 Signal heads should have adequate mechanical strength
alterable only by signal maintenance personnel. High reliability and durability to withstand the conditions of installation and
is also achieved by removing as many moving parts as possible normal use operations. It should be dust proof and
and by using inbuilt self-monitoring and automatic fault weatherproof against corrosion and action of direct sunlight
reporting capabilities. without significant deterioration. The signal head assembly
together with the attachment of a backing board should be
9.4 Essential features of this type of controllers are: capable of withstanding wind velocities up to 160 Km/hour in
any direction and temperatures over a range of -25°C to 70°C.
- ` intelligence' and flexibility inherent to microprocessor
systems offers the possibility of easy modification as 9.15 Normally, the signal heads will be made from
requirements change. polycarbonate or cast aluminium.

- capability of operating in any of the following modes: 9.16 High intensity lamps should be used for enhancing
visibility.
9.5 Hurry Call. This mode maybe requested by the local
detector or by switch input to allow priority to be given to a INDUCTIVE LOOP DETECTORS
particular movement for the passage of emergency vehicles
such as fire engines or rapid transit vehicles. 9.17 The basic system consists of a loop of wire (typically 2
or 3 turns) buried approximately 5mm below the road surface.
9.6 Manual. This mode will allow the controller to operate The ends of the loop are returned to the vehicle detector,
under emergency manual control. usually housed some distance away in the controller cabinet. A
small electric current is passed through the loop and this causes
9.7 Local Coordinate Mode. Multi-plan operation in this an electric field to be built up.
mode shall be possible in accordance with plans and timetables
stored in the memory. 9.18 A change in the inductance of the loop occurs when
a vehicle is positioned directly over it - or is passing over the
9.8 Local Isolated Vehicle Actuated Mode. The con-troller loop. The change in inductance is sensed by the unit and a
in this mode may operate with a mixture of fixed time, demand signal change indicates the presence of a vehicle.
dependent or vehicle actuated phases.
9.19 Provided that loop detectors are properly installed, tested
9.9 Amber flashing on all approaches. and maintained, they will work reliably and offer efficient and
accurate detection as well as providing congestion, counting
- capability of linking to a pedestrian controller for and vehicle classification information.
coordinated control.

32
ASSOCIATED ELECTRICAL WORKS

9.20 These works include

Laying of aspect cables, linking cables (or telecom-

munications cables in an ATC area).

necessary cable jointing work. signal wiring and

connection work.

TRAFFIC SIGNAL CONTROLLER

CIVIL WORKS

9.21 Again, the specifications of civil works vary from

country to country, however, it is recommended that all
cabling should be underground using ducts as shown in
figure 25. Inspection boxes, normally 60cm on each side
should be built every 50m (maximum). In developing
cities the pavement is often dug up for new services, so to
avoid the severance of cables, the ducts should be
protected by a plastic strip in bright colours (normally
yellow with black
identification).

Figure 25 Cross section of cable duct

Plate 3. Microprocessor controller suitable for

pole installation at simple junctions

33
 10. GLOSSARY Early cutoff overlap

A short glossary of terms used in this volume in alphabetical Condition in which one or more traffic streams are permitted to
order is given below: move after the stoppage of one or more other traffic streams,
which during the preceding stage had been permitted to move
All red period with them.

Period when red signals are shown to all approaches simul- Effective green period
taneously, usually of short duration to allow vehicles to clear
the intersection. The period in the green and amber periods throughout which flow
could take place at saturation flow levels.
Audible signals
Extension
Signals in the form of pulsed tones provided for the benefit of
visually handicapped pedestrians. A request for the continuation of the green signal to a
predetermined maximum, made by a vehicle which, when the
request is made, has the right of way.
Area Traffic Control (ATC)

Also called Urban Traffic Control (UTC), this is the cen- Fixed-time traffic signals
tralized control of traffic signals on an area wide basis by
means of computer. Traffic signalling equipment in which the stages and their
duration in each cycle are preset to suit predictable traffic
Backing board conditions.

The Background board is usually coloured black with a white Flow factor
border which is placed behind signal lanterns to make them
more visible against a bright sky or other street or shop lights. The flow factor or `y' value of an approach is the ratio of the
design flow to the saturation flow of the particular approach.
Cycle time
Green filter arrow
A cycle is a complete series of stages during which all traffic
movements are served in turn. The cycle time is the sum of An additional green arrow mounted by the side of the three light
the stage times. display to indicate early movement in the direction of the arrow
and is terminated by a full green light signal. It is normally used
for early discharge of left turners ahead of other movements at
Degree of saturation the same approach.
The degree of saturation at an approach is the ratio of the
design flow to the actual capacity of a particular approach, Green split
weighted by the amount of green the approach receives in a
cycle. The ratio of green time allocated to each of the conflicting phases
in a signal sequence.
Delay
Grid lock
Traffic Delay is the lost time by vehicles due to traffic
`friction', congestion or control devices. The state in which downstream traffic completely blocks the
junction, which then forms queues that blocks other junctions,
Demand and so on until the whole area is blocked with
stopped traffic.
A request for right of way for traffic on a phase which has no
right of way when the request is made. The demands normally Indicative green arrow
being stored in the controller and served in a prearranged
order. An additional green arrow mounted on the right of the three light
display of the secondary signal only, to indicate the early cutoff
Demand - dependent stage/phase of an opposing flow.

A demand-dependent stage or phase is one which appears Intergreen period

only on demand from a vehicle detector or pedestrian push-
button. i.e. it is skippable. The period between the end of the green display on one stage and
the start of the green display on the next stage is known as the
Detector intergreen period.

A device to detect the presence or passage of a vehicle in the Late start overlap
roadway.
Condition in which one or more traffic streams are permitted to
move before the start of one or more of the traffic streams which,
during the subsequent stage are permitted to all run together.

34
Lost time Red-running

Lost time in the green and amber periods is the wastage time The act of disobeying, consciously or not, the red signal
during which no flow takes place. Total lost time per cycle is requiring vehicles on a determined approach to stop and remain
the sum of these lost times for the critical phases plus other lost stationary.
times due to red-amber periods, all red periods and pedestrian
green and flashing green times. Right of way

Maximum green running period The condition which applies when a green signal is displayed
to traffic at the approach thereby permitting that traffic to
The maximum time that a green signal can run after a demand proceed.
has been made by traffic on another phase.
Saturation flow
Minimum cycle time
The maximum flow which could be obtained if 100 percent
The minimum cycle time which is just sufficient to pass the green time was awarded to a particular approach.
traffic.
Semi-vehicle-actuated signals
Minimum green running period
Modified vehicle-actuated signals whereby detectors are
The duration of the green signal following the extinction of a installed on the side roads only. Some semi-vehicle-actuated
red/amber signal during which no change of signal lights can signals also operate one or more demand-dependent stages
occur. within a fixed cycle time.

Offset Stage

The time difference or interval in seconds between the start of A condition of the signal lights which permits a particular
the green indication at one intersection as related to the start of movement of traffic. Stages usually, but not always contain a
the green interval at another intersection from a synchronized green period. They are arranged to follow each other in a
system time base. predetermined order but .stages can be skipped, if not
demanded, to reduce delay.
Optimum cycle time
Spillback
The theoretical cycle time for attaining minimum vehicle delay.
Queue from a downstream junction which affects the traffic
Passenger car units flow at the junction being examined.

Passenger car units for a given type of vehicle are expressed in Traffic signals
terms of the number of moving passenger cars it is equivalent
to, based on headway and delay characteristics. A system of different coloured lights, including arrow-shaped
lights, for controlling conflicting streams of
PCU factor traffic and pedestrians.

An average pcu value derived for the convenience of signal Traffic signal controller
calculation to convert unclassified (by type) vehicle counts
from vehicles per hour units to pcu per hour units. The electronic control equipment which activates the signal
phases at an intersection.
Phase
Vehicle-actuated (V.A.) signals
The sequence of conditions applied to one or more streams of
traffic which during the cycle, receive identical signal light Traffic signalling equipment in which the duration of the green
conditions. Two or more phases may overlap in time. A series time and cycle length varies in relation to the traffic flow on its
of phases is usually arranged in a predetermined order but some approaches.
phases can be omitted if required.
Vehicle extension period
Practical cycle time
A vehicle extension period is the additional duration of the
The cycle time at which the traffic signal installation will be green signal which is secured by the actuation of a detector.
loaded to 90 per cent of its capacity.
Transyt
Reserve capacity
An abbreviation of `Traffic Network Study Tool', a program
A measure of the spare capacity of a signal controlled junction, developed by the Transport and Road Research Laboratory for
expressed in terms of the percentage of the current total flow optimizing signal timings.
factor value (Y) which will be available for further increase of
traffic flows.

35
 11. REFERENCES Vincent, R.A. and Pierce, J.C. MOVA: Traffic Responsive,
Self-Optimising Signal Control for Isolated Intersections.
Burrow, I.J. OSCADY: A Computer Program to Model TRRL Research Report 170. Transport and Road Research
Capacities, Queues and Delays at Isolated Junctions. TRRL Laboratory. 1988.
Research Report 105. Transport and Research Laboratory.
1987. Vincent, R.A, Mitche, A.I. and Robertson, D.I. User Guide to
TRANSYT Version 8. TRRL Report LR888. Transport and
Cannell, A. E. R. and Kaestner, C. Some Aspects of Area Road Research Laboratory. 1980.
Traffic Control in Semi-Developed Countries. Traffic
Engineering acrd Control. 1983. Walker, J.S, Jacobs, G.D. Gardner, G.and Kunchit Phiu-Nual.
The Development of Traffic Management Policies in Bangkok.
Companhia de Engenharia de Transito -CET. Semaforos, Paper presented to CODATU Conference. Jakarta. 1988.
Aspectos de Seguranqa. Sao Paulo. 1990.
Webster and Cobbe. Traffic Signals. Road Research Technical
Department of Transport. Junction Layout for Control by Paper No. 56. H.M.S.O. London, 1966.
Traffic Signals. Highway, Safety and Traffic Advice Note TA
18/81. Department of Transport. 1992. Willumsen, L.G. and Coeymans, J.E. Research into the Value
of Area Traffic Control Techniques in a Developing Contry.
Fouracre, P. R. and Gardner,G. Traffic Signals in Developing TRRL Contractor Report 99. Transport and Road Research
Cities, Overseas Unit Working Paper. Transport and Road Laboratory. 1988.
Research Laboratory. 1990.

Gardner, G, Fouracre, P.R. and Jacobs, G.D. Traffic Man-

agement. Overseas Unit Information Note. Transport and Road
research Laboratory. 1990.

Hunt, P.B, Robertson, D.I, Bretherton, R.D. and Winton, R.I.

SCOOT: A Traffic Responsive Method of Coordinating
Signals. TRRL Report 1014. Transport and Road Research
Laboratory. 1981.

Hurdle, V.F. Signalized Intersection Delay Models- A Primer

for the Uninitiated. Transportation Research Record 971.
Transportation Research Board. Washington, DC. 1984.

Institution of Highways and Transportation/ Department of

Transport. Roads and Traffic in Urban Areas. HMSO, London.
1987.

Jourdain, S. Urban Intersection Control. The Book Guild Ltd.

Lewes, Sussex. 1992.

Robertson, D.I. and Hunt, P.B. A Method of Estimating the

Benefits of Co-ordinating Signals by TRANSYT and SCOOT.
Traffic Engineering acrd Control 1982.

Robertson, D. I. Transyt: A Traffic Network Study Tool. RRL

Report 253. Road Research Laboratory, Crowthorne.1969.

Robertson, D 1, A Method of Estimating the Benefits of

Coordinating Signals by TRANSYT and SCOOT. Traffic
Engineering and Control. London. 1982.

TRL. Urban Road Trafic Surveys. Overseas Road Note 11.

Overseas Centre, Transport Research Laboratory. 1993.

TRL. Design Guidelines for Busway Transit. Overseas Road

Note 12. Overseas Centre, Transport Research Laboratory.
1993.

36
OS-E
ISSN 0951-8987

The use of traffic signals in developing cities ORN13