ISTTT20 Tutorials

Traffic Management
and
Control

A. Traffic Signal Control – Introduction

Prof. Nathan H. Gartner

University of Massachusetts
Lowell, MA, USA

© NHGartner,UML, July 2013

Module Objectives:

 Describe the basic operation principles of traffic signal systems

 Present the principal warrants for traffic signals
 Explain fixed-time, and traffic-responsive operation
 Describe performance measures for isolated intersections
 Develop optimal signal timings for isolated intersections

Students should have received prior instruction in the following areas:

Traffic Flow Principles and Queuing Models

© NHGartner,UML, July 2013

Overview

•Traffic signal control, including sensors,

communications, and advanced timing
algorithms, is a dynamic, continually
evolving form of transportation
management.

•While new technology has resulted in

improved system components and tools
for traffic control signal operations, these
components must be properly designed,
installed, operated, and maintained if
their full potential is to be realized.
Purpose of Traffic Control Signals
The primary function of Traffic Control Signals is to assign
the right-of-way at intersecting streets or highways where,
without such control, a continual flow of vehicles on one
roadway would cause excessive delay to vehicles and/or
pedestrians waiting on the other roadway.

Freeway Ramp Control Signal are a special application of

traffic control signals installed on freeway entrance ramps
to limit, or “meter,” the amount of traffic entering the
freeway.
 Required steps in design:

•investigating the need for a traffic signal

•determining the operational requirements
•translating these requirements into traffic control equipment
requirements
•determining optimum operation of the traffic signal and
•operating and maintaining the traffic control signal over its expected
life.

The information presented in this module builds on procedures documented in

the ITE Transportation and Traffic Engineering Handbook as well as the
standards and warrants provided in the Manual on Uniform Traffic Control
Devices (MUTCD).
Basic Types of Control
Traffic control signals are usually described as either pre-timed or traffic
actuated. Each type may be used in either an independent (isolated) or
interconnected (system) application.

Pre-timed control: The electronic control circuits provide a repetitive

cycle and split (cycle division among the conflicting movements) timing.
The timing is repeated over and over regardless of the presence or
absence of traffic demand. When operating as part of a system, adjacent
intersections operate on the same cycle length and have fixed offsets
(relationship of beginning of main street green displays). There may be
multiple patterns of cycle lengths, offsets and splits.
Basic Types of Control –(cont'd)
 Actuated control: The timing is varied for some or all controlled
conflicting movements dependent upon vehicular and/or pedestrian
demand. Demand is determined from detectors placed in or near the
roadway of pedestrian crossing. When all controlled conflicting
movements are timed relative to demand, the control is termed “full-
actuated.” When only secondary movements vary with demand, the
control is termed “semi-actuated.”

 Full-actuated control: It is generally used for an isolated application. An

isolated application is one where the traffic control signal operates
independent of any other traffic control signal. Conversely, a system or
interconnected application means that a given traffic control signal’s
operation is related to (coordinated with) one or more other traffic control
signal locations. In system applications, a fully actuated controller must
operate as a semi-actuated controller because it must operate on the
same cycle length as all the other controllers that are interconnected with
it. In systems, pre-timed and actuated control may be mixed.
 ISTTT20 Tutorials

Traffic Management
and
Control

B. Needs Assessments

Prof. Nathan H. Gartner

University of Massachusetts
Lowell, MA, USA

© NHGartner,UML, July 2013

Section B: Needs Assessment -
Determining the Need for Traffic Signal Control

The first and basic question that must be

addressed is whether or not traffic signalization is
needed. Since traffic signals are the most
restrictive traffic control devices, they should be
used only where the less restrictive signs or
markings do not provide the necessary level of
control.
Required Studies

A comprehensive investigation of traffic conditions and physical

characteristics of the location is required to determine the necessity for a
signal installation. The engineering study should include:

1. The total number of vehicles entering the intersection in each hour from
each approach during 16 consecutive hours of a representative day. The 16
hours selected should contain the greatest percentage of the 24-hour traffic.

2. Vehicular volumes for each traffic movement from each approach, classified
by vehicle type during AM and PM 2-hour peak.

3. Pedestrian volume counts on each crosswalk during the same periods as the
vehicular counts .
Required Studies – (cont'd)

4. Information about nearby facilities and activity centers that serve the
young, elderly, and/or persons with disabilities.

5. The 85-percentile speed of all vehicles on the uncontrolled approaches to

the location.

6. A conditions diagram showing details of the physical layout, including such

features as intersection geometrics, channelization, grades, sight-distance
restrictions, bus stops and routings, parking conditions, etc.

7. A collision diagram showing accident experience by type, location,

direction of movement, severity, time of day, date and day of week for at
least one year.
Required Studies – (cont'd)

The following data are also desirable for a more thorough understanding of
the operation of the intersection and may be obtained when volume data
by approach is being collected (the second step described on the previous
page):

1. Vehicle-hours of stopped time delay determined separately for each

approach.
2. The number and distribution of acceptable gaps in vehicular traffic on
the major street for entrance from the minor street.
3. The posted or statutory speed limit or the 85th-percentile speed on
controlled approaches at a point near to the intersection but
unaffected by the control.
4. Pedestrian delay time for at least two 30-minute peak pedestrian delay
periods of an average weekday or like periods of a Saturday or Sunday.
Required Studies – (cont'd)

Adequate roadway capacity at a signalized intersection is

desirable.
Widening of both the major street and the minor street may
be warranted to reduce the delays caused by assignment of
right-of-way at intersections controlled by traffic signals.
Widening of the minor street is often beneficial to operation
on the major street because it reduces the green time that
must be assigned to minor street traffic.
Warrants for Traffic Signal Installation

Traffic control signals should not be installed unless one or more of the
signal warrants in the MUTCD are met:

 Warrant 1 - Condition A, Minimum Vehicular Volume and

Condition B, Interruption of Continuous Traffic
 Warrant 2 - Four-Hour Vehicular Volume
 Warrant 3 - Peak Hour
 Warrant 4 - Pedestrian Volume
 Warrant 5 - School Crossings
 Warrant 6 - Coordinated Signal System
 Warrant 7 - Crash Experience
 Warrant 8 - Roadway Network
Warrant 1: Eight-Hour
VehicularVolume
Warrant 1: Eight-Hour Vehicular Volume - (cont'd)
Warrant 1: Eight-Hour Vehicular Volume - (cont'd)
Warrant 1: Eight-Hour Vehicular Volume - (cont'd)
Warrant 1: Eight-Hour Vehicular Volume - (cont'd)
Warrant 2: Four-Hour Vehicular Volume
Warrant 2: Four-Hour Vehicular Volume - (cont'd)
Warrant 3: Peak Hour
Warrant 3: Peak Hour – (cont'd)
Warrant 3: Peak Hour – (cont'd)
Warrant 4: Pedestrian Volume
Warrant 4: Pedestrian Volume - (cont'd)
Warrant 5: School Crossing
Warrant 6: Coordinated Signal System
Warrant 7: Crash Experience
Warrant 8: Roadway Network
Warrant 8: Roadway Network - (cont'd)
 ISTTT20 Tutorials

Traffic Management
and
Control

C. Operational Requirements

Prof. Nathan H. Gartner

University of Massachusetts
Lowell, MA, USA

© NHGartner,UML, July 2013

Section C: Operational Requirements

After it has been established that a signal is warranted at a

particular location (or locations), the next major decision in the
design process involves determining the most appropriate
method of control. Decision to be made include:

 Controller phasing

 Pre-timed or actuated operation

 Interconnection considerations
Phasing Elements

A signal phase may be defined as that part of the cycle length

allocated to a traffic movement receiving the right of way or to
any combination of traffic movements receiving the right of way
simultaneously.

A traffic movement is a single vehicular movement, a single

pedestrian movement, or a combination of vehicular and
pedestrian movements.

The sum of all traffic phases is equal to the cycle length.

Example: 3-phase controller sequence
Phasing Elements (cont'd)

Phases are commonly added for protecting left turns. The basic sequences
which accommodate left-turn movements include:
Heaviest left turn protected. This is a "lead left" in which the left-turning
vehicles from only one approach are protected and move on an arrow
indication preceding the opposing through movement; or a "lag left" in
which the protected left turn follows the trough movement phase.

Both left turns protected-no overlap. When the opposing left turns move
simultaneously followed by the through movements, it is termed "lead dual
left." If the left turns follow the through movement it is called a "lag dual
left."
Phasing Elements (cont'd)

Both left turns protected with overlap. In this operation, opposing left turns
start simultaneously. When one terminates, the through movement in the
same direction is started. When the extending left is terminated, the other
through movement is started. When this type of phasing is used on both
streets, it is termed "quad left phasing.“

Lead lag. This phasing is combined with a leading protected left in one
direction, followed by the through movements, followed by a lag left in the
opposing direction. It is sometimes used in systems to provide a wider two-
way through band.

Directional separation. Each approach obtains exclusive right of way with all
opposing traffic stopped.
Example: 8-phase dual ring controller
Types of Control

The principal types of traffic signal control are pre-timed and

traffic actuated. Each type of control has its unique
advantages and disadvantages.

Pre-timed control assigns the right of way at an intersection

according to a predetermined schedule. The length of the
time interval for each signal indication in the cycle is fixed,
based on historic traffic patterns.
Actuated & Semi-Actuated Control

Actuated control differs from pre-timed in that signal phases

are not of fixed length. Through the use of vehicle detectors,
this type of control assigns the right of way on the basis of
current traffic conditions (demand) within given limitations.

Fully-actuated control requires detectors for all phases with

each phase timed according to preset parameters. Fully-
actuated control is primarily used at the intersection of streets
with approximately equal volumes with sporadic and varying
traffic distribution.
Actuated & Semi-Actuated Control – (cont'd)

Semi-actuated control requires detectors on the minor street

approaches and is especially effective at intersections where
the major street has a relatively uniform flow and the minor
street has low volumes with random peaks. Semi-actuated
control, therefore, only takes “away” time from the major
street movements when there is actually demand on the minor
street.
Choosing a Type of Control

In general practice, the rule of thumb for choosing the type of

intersection control is:
•for predictable traffic demand, use pre-timed (Warrant 1);
•for unpredictable traffic demand, use actuated control (Warrant
2).
Interconnection Considerations

The potential benefits to be derived from coordinated operation

of two or more signalized intersections are directly related to
the “platoon” arrival characteristics at the downstream
intersection.

If approaching vehicles arrive at the stop line as a well-defined

compact platoon, coordinated operation can provide a
significant reduction in stops and delays.

The MUTCD suggests that signals spaced less than ½ miles apart
should be coordinated because the cohesion of the platoon can
be maintained for this distance.
 ISTTT20 Tutorials

Traffic Management
and
Control

D. Single Intersection Models

Prof. Nathan H. Gartner

University of Massachusetts
Lowell, MA, USA

© NHGartner,UML, July 2013

Section D: Single Intersection Models

The material in this section is based on the

methodology developed by F. V. Webster at the
British Road Research Laboratory for estimating
the delays to vehicles at fixed-time traffic signals
and for computing the optimum settings of such
signals.
Section D: Models – (cont'd)

We will begin our discussion of traffic signal

settings by describing two key concepts:
•delay and
•saturation flow rate.
Definition of Delay

The most commonly used measure of effectiveness for

signalized intersections is delay.

Delay to the traffic is computed by counting the number of

vehicles in the queue at fixed intervals of time and multiplying
this number by the value of the time interval.
Saturation Flow

Saturation flow occurs when the

queue discharges at a more or
less constant rate.

The saturation flow rate is

measured in units of vehicles per
hour of greet time per lane.
Effective Green Time
If there is still a queue at the end of the green period, some vehicles will make
use of the yellow interval to cross the intersection. In these circumstances
traffic moves on both green and yellow signals. To simplify analysis, it is
advantageous to use the concept of “effective green time.” The green and
yellow periods together may be considered as an 'effective' green (g) and a
'lost' time (l).

The effective green time is determined such that the product of the effective
green time and the saturation flow rate is equal to the actual number of
vehicles (say, b) discharged from the queue on the average in a saturated
green period (i.e. a green period during which the queue never clears).
Signalized Intersections: Fluid Traffic Model

In this section a continuum (i.e. deterministic) or fluid model is first considered

for which basic measures of queue length and delay are developed. This is a
relatively simple model that does not take into account the variability of traffic
(i.e. differing numbers of vehicles arrive at intersections at different times).

In estimating delay at intersections, traffic flow is considered as consisting of

identical passenger car units (pcu's). A truck, for example, may be considered as
1.5 or 2 pcu's and a turning vehicle may also be assigned some value depending
on the type of maneuver that is made.

For the discussion of this model, the following variables (defined below) will be
used:
Signalized Intersections: Fluid Traffic Model – (cont'd)

Let

c = the cycle time (sec) g = the effective green time (sec)

r = the effective red time (sec)
q = the average arrival rate of traffic on the approach (pcu/sec)
s = the saturation flow on the approach (pcu/sec)
d = the average delay to a pcu on the approach (sec)
Q0 = the overflow (pcu)
λ = g/c (the proportion of the cycle that is effectively green
y = q/s (the ratio of average arrival rate to saturation flow)
x = qc/gs (the ratio of average number of arrivals/cycle to the maximum
number of departures/cycle)

Thus, r + g = c and λx = y. The ratio x is called the degree of saturation of the

approach and y is called the flow ratio of the approach.
Signalized Intersections: Fluid Traffic Model (cont'd)

First, let’s consider an equation to model the cumulative time that is

required for a queue of n stopped vehicles at an intersection to pass the
signal once the green interval begins. Because of lost time due to starting,
the first several vehicles take longer to pass the signal than do vehicles back
in the queue.

According to a simple model of traffic flow (in this case, developed by

Greenshields the successive time intervals for vehicles departing from a
queue to cross the signal stop line are : 3.8, 3.1, 2.7, 2.4, 2.2, 2.1, 2.1....
seconds. Thus, once the fifth car in the queue arrives, all subsequent
vehicles will arrive at 2.1 second intervals. The cumulative time for a queue
of n stopped vehicles to pass a signal can then be given by

Cumulative time = 14.2+2.1(n-5) sec for n≥5

Signalized Intersections: Fluid Traffic Model (cont'd)

If the lost time due to starting had been ignored, and all of the vehicles
departed at the saturation flow rate s = 1/2.1[veh/sec] = 1714 [veh/hr],
the first five vehicles would have required only 10.5 seconds to clear the
intersection. Therefore, since these vehicles actually required 14.2
seconds to clear, one will see that the effective green time is the signal
green time less 3.7 seconds (i.e. 14.2 – 10.5).
Signalized Intersections: Fluid Traffic Model (cont'd)

Before exploring equations to model

delay, consider graphically arrivals
and departures of vehicles at a
signalized intersection. Delay can be
calculated by illustrating the meaning
of arrival time and departure time for
a pcu on an approach. They are
demonstrated by reference to Figure
D-3, in which distance-time curves
are plotted for each of four vehicles.
Signalized Intersections: Fluid Traffic Model (cont'd)

The line AB on the far left represents the passage of an un-delayed vehicle,
where the horizontal line PQ represents the stop line at which the first
vehicle waits when there is a queue. CDEF represents the trajectory of the
first vehicle that is delayed by a signal. The straight portions of CD and EF are
parallel to AB and projected to meet PQ and X and Y so that the length XY is
the delay to the first vehicle.

In other words, if the first vehicle had not been delayed by the signal, it
would have reached the stop line at time X. However, because it had to slow
down/stop (the period between D and E), given the point it is at at E, it
would have crossed the stop line at Y if there had been no delay. Therefore,
the total delay for the first vehicle is XY. Similarly, X'Y' and X"Y" represent the
delays for the next two vehicles.
Fluid Model for Pre-timed Signal
Another representation of a fluid model of
traffic flow at a signalized intersection is
shown in Figure D-4. The vertical axis
represents the number of vehicle arrivals
at the stop line and the horizontal axis
represents the time t. The figure
illustrates the behavior when the capacity
of the green interval can fully
accommodate (i.e. serve) the number of
arrivals during the green + red time - i.e.,
an under-saturated period.

The top figure illustrates the queue length

as a function of time (t0 is the queue
clearance time, measured from the start
of green ). The bottom figure illustrates
the cumulative arrival and departure
functions.
Fluid Model for Pre-timed Signal – (cont'd)

In Figure D-4 the vertical distance ca represents the number of vehicles that
have accumulated since the signal entered the red phase at time c. At time r,
the red interval is over, and the queue begins to clear. This is represented by
the line forming the “right-hand side” of the triangle. Once the queue
clearance line meets the queue formation line (the “left-hand side” of the
triangle) the queue is completely cleared.

The horizontal distance ab represents the total time from arrival to

departure for any given vehicle. The area of the triangle in the top figure
represents the total vehicle delay. This is identical to the shaded area in the
bottom figure.
Fluid Model for Pre-timed Signal (cont'd)

The following measures of queue behavior are developed:

1. Time after start of green that queue is dissipated (t0)

2. Proportion of cycle with queue (Pq)

3. Proportion of vehicles stopped (Ps)

4. Maximum number of vehicles in queue (Q m)

5. Average number of vehicles in queue (Q)

6. Total vehicle-time of delay per cycle (D)

7. Average individual vehicle delay (d)

8. Maximum individual vehicular delay (dm)

Fluid Model for Pre-timed Signal (cont'd)

On the next pages, we will present and describe a number of equations that
can be used to analyze delay at an intersection based on the fluid model. The
equations have been developed based on the graphs presented on the
previous two pages. Students are encouraged to refer back to these graphs to
better understand the foundation of the equations:

1. For any given cycle, as with any physical system, “what goes in, must come
out.” In other words, once time t0 (the time after green at which the queue
is completely cleared) is reached, the total number of vehicles that arrived
and waited in the queue must be equal to the total number of vehicles that
have left the queue.
Fluid Model for Pre-timed Signal (cont'd)
In the equation below, remember that q is the arrival flow rate, and s is the
saturation flow rate (i.e. capacity of the intersection). Thus, the variable y is
the ratio of the arrival rate to saturation flow rate. As this ratio becomes
closer and closer to 1, the time to clear the queue, t0, becomes larger and
larger.

Equation D.1

The total number of arrivals is

q(r + t0 ) = st0
Letting y = q/s, we obtain
t0 = yr/(1 - y)
Fluid Model for Pre-timed Signal (cont'd)

2. The proportion of vehicles delayed by the intersection control is equal to

queue time (represented by the red interval plus t0) divided by the total cycle
length.

Equation D.2

Pq = (r + t0)/c
Fluid Model for Pre-timed Signal (cont'd)

3. The proportion of vehicles stopped is equal to vehicles stopped/total

vehicles per cycle.

Equation D.3

Ps = q(r + t0) / q(r + g) = t0 / (yC)

Fluid Model for Pre-timed Signal (cont'd)

4. The maximum number of vehicles in queue can be seen by examining Figure D-4. As
seen in the figure, it is the height of the triangle at t = r, i.e., at the end of the red
period. This is logical in that the maximum queue results at the end of the time that
vehicles (at the average arrival rate, q) have been arriving during the red phase:

Equation D.4

Qm = qr
Fluid Model for Pre-timed Signal (cont'd)
5. The average number of vehicles in the queue at any given time can be
calculated by dividing the total number of vehicle-seconds of delay during
the cycle length (represented by the area of the triangle in Figure D.4),
divided by the total cycle length, C (or r + g). Recall that the area of a triangle
is half of its base length time its height – this results in the expression for
total delay of [(qr/2)r + (qr/2) t0]:

Equation D.5
Fluid Model for Pre-timed Signal (cont'd)

6. As described on the previous page, the total vehicle-time of delay during

the cycle is equal to the area of the triangle in Figure D-4 (which, in terms of
Calculus, is the integral of the queue-length curve). In equation D.6 below,
we simply substitute the variable Y (ratio of average arrival rate to
saturation flow rate) to derive an equation that models total delay based on
the average arrival rate, red interval, and Y.

Equation D.6
Fluid Model for Pre-timed Signal (cont'd)
7. The average individual delay, d, can be calculated by dividing the total
vehicle-time delay (derived on the previous page) by the number of vehicles
that arrive during the cycle, qc.

Equation D.7
Fluid Model for Pre-timed Signal (cont'd)

8. Finally, the maximum individual vehicular delay can be seen from Figure
D-4 to be simply equal to the total red interval. This corresponds to the
vehicle that just arrives at the intersection as the signal turns to red.

Equation D.8
Fluid Model for Pre-timed Signal (cont'd)

If the possible departures during the cycle (i.e., the capacity of the intersection
– the saturation flow rate times the total green interval - sg) are less than the
total number of arriving vehicles during the entire cycle, qc, the queue grows
with each successive cycle.

In other words, the queue that formed during the red interval is not cleared
during the green interval – and some vehicles find themselves delayed by more
than one cycle of the intersection. This situation is described as oversaturation.
In the case of oversaturation – the previous equations are not applicable and
cannot be used for analysis.
Fluid Model for Pre-timed Signal (cont'd)
Example:
Consider the following example of the behavior of a queue at a signalized
intersection. Assume that the green phase g is 40 sec, the red phase r is 20 sec,
the discharge rate s is 1,200 vehicles/hr, and there are two input flow rates q1 =
600 vph; q2 = 800 vph. In the first case, the capacity of the green interval
exceeds the number of arrivals during the green + red time (the undersaturated
case). In the second case, the discharge during the green phase is equal to the
arrivals during the green + red period (the saturated case). The results are given
in Table D-1 (below).
Average Delay Per Vehicle – WEBSTER Model

As stated before, the equations developed and presented up to this point are
only applicable for undersaturated deterministic conditions; i.e., when volume
is below capacity and the arrival pattern is known apriorily. Clearly, in many
situations, this is not appropriate for use in signal analysis.

Webster developed a more complex model of average delay per vehicle that
assumes random arrival of vehicles (in this case, based on the Poisson
distribution). Although the v/c ratio still must be less than 1.0 for the entire
analysis period, there are possible overflows due to the random nature of the
arrivals. We will explore this model in more depth on the following pages.
Average Delay Per Vehicle – WEBSTER Model – (cont'd)
Average Delay Per Vehicle – WEBSTER Model – (cont'd)
To simplify the equation above, we rewrite the equation as a function of the
cycle length, c, and the arrival flow rate, q. Then, the equation uses two
terms – A and B – these are functions of the green time ratio and the ratio of
maximum flow to capacity. Remember, equation D.10 is exactly the same as
D.9 presented above – it is just written in a more simple manner.

Equation D.10
Average Delay Per Vehicle – WEBSTER Model – (cont'd)

The first term of the delay equation – cA – models delay when vehicles are
arriving at a uniform rate (i.e. no randomness in the arrival of vehicles). In fact,
this term is identical to the simpler model presented in Equation D.7. At low
flow arrival flow rates, this first term does a very good job of modeling delay.
However, at higher flow rates when randomness becomes a larger factor, the
term alone cannot effectively model delay.
Average Delay Per Vehicle – WEBSTER Model – (cont'd)

The second term of Equation D.9 is often referred to as the “overflow delay”
component of the model. It makes some allowance for the random nature of
the arrivals. It is an expression for the delay experienced by vehicles arriving
randomly in time at a 'bottleneck', queuing up, and leaving at constant
intervals. In queuing theory terminology it is the average delay of a M/D/1
queue. The addition of the second term improves the effectiveness of the delay
model – but often overestimates delay. To avoid the need to use an empirical
correction factor C, a suitable approximation can be obtained by reducing the
delay by a percent P in one of the following ways:

Equation D.11
Average Delay Per Vehicle – WEBSTER Model: Example

The next pages present a numerical example of applying the delay model. In
this example, the average flow on a particular approach of an intersection is
600 vehicles per hour, and the signal settings are 30 seconds green, 4 seconds
yellow and 60 seconds cycle time. It is observed that on the average 15.0
vehicles are discharged in a fully saturated green period.

If we assume that starting delays, etc. are responsible for 2 seconds of each
green-plus-amber period then 15 vehicles are discharged in an effective green
time of 30 seconds, i.e. s = 1800 vehicles per hour. It is assumed that 2
seconds of the yellow time are also included as part of the effective green.
Based on this given information, we can compute the green interval portion
of the cycle time (λ), the capacity flow (K) and the degree of saturation (x).
Average Delay Per Vehicle – WEBSTER Model:
Example – (cont'd)
Average Delay Per Vehicle – WEBSTER Model: Example
– (cont'd)

From the delay model equations, we can calculate that A = 0.187 and B =
0.667. Finally, assume that in this case, C = 9 percent of the total delay as
estimated by first two terms. Using Equation D.10, we estimate average
vehicular delay to be:
Average Delay Per Vehicle – WEBSTER Model: Example
– (cont'd)

Alternatively, we can use the approximations according to Equation D.11:

 ISTTT20 Tutorials

Traffic Management
and
Control

E. Traffic Signal Settings

Prof. Nathan H. Gartner

University of Massachusetts
Lowell, MA, USA

© NHGartner,UML, July 2013

 Optimum Settings of Fixed-Time Signals

The objective in setting signal timings for a fixed-time signal is

to minimize overall vehicular delay. This is in addition to the
need to meet the warrants described earlier.
Green Times Determination

The ITE Transportation and Traffic Engineering Handbook [1]

states that the least delay to traffic at an intersection is obtained
when the green periods of the phases are set in proportion to
the corresponding ratios of flow to saturation flow, on the
dominant (or critical) approaches.

This division of the cycle time makes the capacity of the phases
proportional to the average flows of the phases.
Green Times Determination
- Example CYCLE TIME
(seconds)
20

Typical results are shown

AVERAGE DELAY PER VEHICLE - seconds

in Fig. D-6, where the 80

best ratio of the y values 15

60

is approximately 2.0. It 35

can be seen that the best 40

ratios of the effective

10
green times is between
1.88 and 2.17 over a N

qN=600 q - Flow: vehicles per hour

range of cycle times of 35 W E

s - Saturation flow: 1800 vehicles
per hour on all approaches

to 80 seconds. qW =300 qE=300 L - Lost time per cycle: 10 seconds

g - Effective green time
qS=600

0 1.0 2.0 3.0 4.0

qN-S/qE-W

Fig. D-6 Effect on delay of the variation of the ratio

of green periods
Optimum Cycle Time
The challenge in this case is to determine the value of cycle
time which gives the least delay of all traffic using the
intersection.

Webster developed a relatively simple expression to determine

optimal cycle length – based on total lost time and the sum of
the y values at all approaches (recall that y is the ratio of the
arrival flow rate to the saturation flow rate at an approach):

1.5L  5
co  sec
1 Y

where Y is the sum of the y values for all phases and refers to the intersection
as a whole and L is the total lost time per cycle in seconds.
Cycle Time (cont'd)
The lost time value in the optimal cycle
length equation can be expressed by

L = nl + R

where:

n is the number of phases

l is the average lost time per phase

(excluding any all-red periods)

R is the total time during each cycle when

all signals display red simultaneously.
Cycle Time (cont'd)
Cycle Time (cont'd)

When traffic is of a truly random character, the minimum cycle

time will result in nearly infinite delay.

The optimal cycle length (i.e. the cycle length that results in the
minimum amount of total delay at the intersection) is roughly
two times the minimum cycle length, therefore:
Guidelines for Selection of Cycle Time for Fixed-Time Signals

A challenge in designing signal settings for a fixed-time signal lies in the fact
that, in most cases, arrival volumes vary throughout the day. For example,
traffic is often heavier during peak periods than mid-day and at night.
Remember that the approach we have discussed so far relies on developing
settings for particular arrival volume levels. Ideally, the signal will have different
settings that are designed for different times of day (with the corresponding
different volume levels). In some cases, vehicle-actuated signals can also be
used to account for this.

However, when these approaches are not feasible, another compromise is

possible.
Guidelines for Selection of Cycle Time for Fixed-Time Signals –
(cont’d)

It can be seen that the delay for cycle times within the range of ¾ to 1½ times
the optimum cycle length value is never more than 10 to 20 percent greater
than delay for the optimum cycle. Based on this observation, for a single
setting of fixed-time signals the simple approximate method outlined below
may be used.

i. Calculate the optimum cycle for each hour of the day when the traffic flow
is medium or heavy, e.g. between the hours of 8 a.m. and 7 p.m. and
average over the day.

ii. Evaluate three-quarters (¾) of the optimum cycle calculated for the
heaviest peak hour.

iii. Select whichever is greater for the cycle time.

Guidelines for Selection of Green Times for Fixed-Time Signals –

It is suggested as a reasonable procedure that the division of the

available green time during the cycle (c0 - L) should be in
proportion to the average y values for peak periods .

g1 ( y1 ) PEAK

g 2 ( y2 ) PEAK
Signal Settings: Example

Example: determine settings for a 2-phase, 4-arm intersection.

First, the following data is collected in the field:

North South East West

Flow (q ) vehicle per hour … 600 450 900 750
Saturation flow (s) " " " … 2400 2000 3000 3000
Ratio (q/s) … …. … …. 0.25 0.225 0.3 0.25
Y values … … … 0.25 0.3
Lost time
Starting delays …. … … … … …. 3… seconds per phase
All-red periods … … … … .. 3 seconds at each change of right of way
Signal Settings: Example (cont’d)
Assuming a 4-second yellow interval, the total lost time per phase is 3+4(.5)
= 5 sec. For the two phases it is 10 seconds (see the figure below).
Signal Settings: Example (cont’d)
The total lost time for the cycle, L, including all-red periods, is
therefore 16 seconds.

The optimal cycle length for the intersection is calculated as

follows:

1.5(16)  5 29
c0    64 sec
1  0.250  0.300 0.450
 Signal Settings: Example (cont’d)

Now, given the total cycle time of 64 seconds, it is necessary to set green
intervals for both phases. The y value for the North-South phase is 0.25, while it
is 0.3 for the East-West phase. We remove the all-red time periods and yellow
periods from the total cycle length to determine “available” total green: 64
seconds – 6 seconds – 8 seconds = 50 seconds.

Therefore:

Green Interval for North-South: 50 seconds(0.25/[0.25+0.3]) = 23 seconds

Green Interval for East-West: 50 seconds(0.30/[0.25+0.3]) = 27 seconds