2

Types of Traffic
Flow
and
Traffic Stream
Parameters
Outline

2. Introduction Concepts in Transportation Engineering

2.1 Introduction
2.2 Types of Flow
2.3 Major Traffic Variables
2.4 Other Traffic Variables
2.5 Models of Traffic Flow
2.6 Sample Computations

Learning Outcomes
1. Describe and estimate traffic characteristics including
a. time mean speed and space mean speed;
b. flow and headway;
c. density and space headway; and
d. occupancy
2. Understand the fundamental relationship between flow, speed and density.
2.1 INTRODUCTION
Describing traffic is considered very difficult due to several factors that cause its
irregularity or unpredictability. These factors could be attributed to several events which could
happen on the road: accidents, stalled vehicles, lane changing or swerving, parking maneuvers,
indiscriminate loading and unloading of public utility vehicles, etc. Without these events, traffic
flow could be expected to be fairly regular and predictable within a day or even within a week.
However, the more frequently these events occur along the road, the harder it is to predict the
traffic condition.
It is during the occurrence of a traffic event when more advanced knowledge of traffic
flow theory, like queuing analysis or shock wave, becomes necessary. These topics will be
introduced at the end of the chapter.
First, let us consider how we can best describe traffic condition when congestion builds
up simply due to continuous increase of traffic density along a road. To illustrate, imagine the
development of traffic along the expressway. In the early morning, drivers can freely choose their
own speeds because there are only very few vehicles on the road. As traffic density increases,
the drivers are constrained to adjust their speeds. Traffic condition becomes very unstable when
the capacity of the highway has already been reached, after which, a stop-and-go condition will
be experienced if the situation worsens, traffic is brought to a standstill - a complete breakdown
of the expressway because of lack of capacity. This condition may last for several minutes or even
hours. However, we can expect a reverse process to occur when traffic flow returns to stable
condition.

2.2 TYPES OF FLOW

Traffic flow is usually classified as either uninterrupted or interrupted. Flow occurring at
long sections of road where vehicles are not required to stop by any cause external to the traffic
stream is called uninterrupted flow. On the other hand, flow occurring at intersections or
driveways where vehicles are required to stop by any cause outside the traffic stream - such as
traffic signs (STOP or YIELD), traffic signal lights, etc. - is called interrupted flow. For each type of
flow, we will discuss the most commonly used traffic variables that best describe them.
2.2.1 MAJOR TRAFFIC VARIABLES
Uninterrupted flow can be described using any of the following traffic variables:
A. Flow rate or volume
B. Speed
C. Density or concentration

Flow rate or volume

Flow rate is defined as the number of vehicles passing a point during a specified period of
time. It is often referred to as volume when measured over an hour. Figure 1 illustrates the
trajectories of vehicles traveling in given section L of a one-way road. Consider a point on the
road (represented by a line 1-1 drawn transversely).

Figure 1 Time-Distance Diagram

If N vehicles (represented by trajectories crossing line 1-1) passed this point in time T,
flow rate q may be expressed as:

𝑛
𝑞= Equation 1
𝑡

Again, if the observation period T is set to one hour, q is called volume and will have a
unit of vehicles per hour. In general, flow rate (or volume, as used loosely) will have units like
vehicles per minute or vehicles per day.

Speed
 Speed is defined as rate of motion in distance per unit time. When describing traffic
stream, two types of speed are used: time men speed and space mean speed.
a. Time mean speed
Also called spot speed, time mean speed is simply the arithmetic mean of the
speeds of vehicles passing a point within a given interval of time. Strictly speaking,
distance or length of road must be known in order to measure speed. However, with the
use of speed radar, spot speed can be measured at a certain point on the road. Also, spot
speed can be reasonably measured if a point is approximately by a short distance, say 15-
50 m of road. This distance is normally called trap length in spot speed studies. With the
use of stop watch, the time it takes for the vehicle to traverse this trap length is measured.
Individual speed of vehicles is then computed by dividing this trap length by this measured
time. Figure 2 shows how spot speed is approximated by the use of trap length, ∆𝑥.

Figure 2 Measuring spot-speed using trap length method

where:
ui = speed of vehicle I, in kilometre per hour
ti = time it takes for vehicle to traverse the trap length
Δx = trap length, in meters

then ui

Knowing the individual speeds of n vehicles observed within time T, the time mean
speed or spot speed of the traffic stream is given by

∑𝑛
𝑖=1(𝑢𝑖 )
𝑢̅𝑡 = Equation 2
𝑛

where
 𝑢̅𝑡 = time-mean speed in unit distance per unit time,
𝑢𝑖 = spot speed (the speed of the vehicle at the designated point on the
highway, as might be obtained using a radar gun) of the ith vehicle, and
n= number of measured vehicle spot speeds.

b. Space mean speed

Space mean speed is used to describe the rate of movement of a traffic stream
within a given section of road. It is the speed based on the average travel time of vehicles
in the stream within the section. It is also called the harmonic mean speed. Consider the
time-distance diagram shown in figure 3. Only the speeds of vehicles intersected by the
line drawn at instant time t are measured. (The speed of a vehicle can actually be
estimated by dividing the length of the road section L by its travel time.)

Figure 3 Time-distance diagram for space men speed

If n vehicles are observed at an instant of time t, the space mean speed is

computed as follows:
𝑛
𝑢𝑖 = 1 Equation 3
∑𝑛
𝑖=1(𝑢 )
𝑖

Density

Density is defined as the number of vehicles in a given length of road at an instant point
in time. Using the same figure 3, the number of vehicles counted at time t divided by the length
of the section L gives a measure of density in that section. Again, if n vehicles are found within
the section L, density k is computed as:
𝑛
𝑘= Equation 4
𝑙

Among these three variables, density proves to be the most difficult or expensive to
observe. Aerial photography is the most commonly used method to get its exact values. On the
other hand, there are various ways of conducting volume and speed studies. It will be seen later
that these variables are actually related to each other and that density can be derived if volume
and speed are known.

2.2.2 OTHER TRAFFIC VARIABLES

There are other variables used to describe traffic flow. These variables, however, are
simply variants of the three variables described previously.

Time headway
Time headway is defined as the time interval between passage of consecutive vehicles at
a specified point on the road with a unit of time per vehicles. Referring again to figure 1 where N
vehicles were counted to pass line 1-1 within the observation time T, let hi be the time headway
of 2 consecutive vehicles. The average headway representing all observed vehicles is estimated
using arithmetic mean. (Note that if there are n vehicles observed, then there will be n - 1
headways that will be measure.)

ℎ𝑡 = ∑𝑛−1
𝑖=1 ℎ𝑖 Equation 5

where
t = duration of time interval,
hi = time headway of the ith vehicle (the elapsed time between the arrivals of vehicles i
and i − 1), and
n = number of measured vehicle time headways at some designated roadway point.

For longer observation period:

𝑡
∑𝑁−1
𝑖=1 ℎ𝑖 ≅ 𝑇 and 𝑛 − 1 ≅ 𝑛, or ℎ𝑡 ≅ Equation 6
𝑛

Therefore, the average time headway and flow rate are related as follows:
1
ℎ𝑡 = 𝑞 Equation 7

Spacing
Spacing is the distance between two vehicles measure from the front bumper of a vehicle
to that of another. Similar to the estimation of time headway, if there are n vehicles within a
given road section L, the sum of (n -1) spacing si will be almost equal to L. Average spacing,
therefore, may be computed as the inverse of density.

1
𝑠=𝑘 Equation 8

Time occupancy
Another useful measure of traffic flow is time occupancy. It can only be measure,
however, if a detector is installed at a specific point on the carriageway. It is defined as the total
time of a detector is occupied divided by the total time of observation. Figure 2.4 shows how a
vehicle is detected using ultrasonic detector.

Figure 4 Vehicle detection using ultrasonic detector

*Assuming that n vehicles were observed during the total time of observation T,
the time occupancy Ot ,is given by

∑𝑛
𝑖=1 𝑡1
𝑂𝑡 = 𝑥100 Equation 9
𝑇

where ti is the detection time of the ith vehicle.

2.3 MODELS OF TRAFFIC FLOW

With the basic relationships among traffic flow, speed, and density formalized, attention
can now be directed toward a more microscopic view of traffic flow. That is, instead of simply
modeling the number of vehicles passing a specified point on a highway in some time interval,
there is considerable analytic value in modeling the time between the arrivals of successive
vehicles (the concept of vehicle time headway presented earlier). The most simplistic approach
to vehicle arrival modeling is to assume that all vehicles are equally or uniformly spaced. This
results in what is termed a deterministic, uniform arrival pattern. Under this assumption, if the
traffic flow is 360 veh/h, the number of vehicles arriving in any 5-minute time interval is 30 and
the headway between all vehicles is 10 seconds (because h will equal 3600/q). However, actual
observations show that such uniformity of traffic flow is not always realistic because some 5-
minute intervals are likely to have more or less traffic flow than other 5-minute intervals. Thus, a
representation of vehicle arrivals that goes beyond the deterministic, uniform assumption is
often warranted.

2.3.1 Poisson Model

Models that account for the nonuniformity of flow are derived by assuming that the
pattern of vehicle arrivals (at a specified point) corresponds to some random process. The
problem then becomes one of selecting a probability distribution that is a reasonable
representation of observed traffic arrival patterns. An example of such a distribution is the
Poisson distribution (the limitations of which will be discussed later), which is expressed as

(𝜆𝑡)𝑛 𝑒 −𝜆𝑡
𝑃(𝑛) = Equation 10
𝑛!

Where
𝑃(𝑛) = probability of having n vehicles arrive in time t,
𝜆 = average vehicle flow or arrival rate in vehicles per unit time,
𝑡 = duration of the time interval over which vehicles are counted, and
𝑒 = base of the natural logarithm (e = 2.718).

The assumption of Poisson vehicle arrivals also implies a distribution of the time intervals
between the arrivals of successive vehicles (time headway). To show this, note that the average
arrival rate is
𝑞
𝜆 = 3600 Equation 11

where
λ = average vehicle arrival rate in veh/s,
q = flow in veh/h, and
3600 = number of seconds per hour.

Substituting Eq. 10 into Eq. 11 gives

𝑛 −𝑞𝑡
𝑞𝑡 ( )
(3600) 𝑒 3600
𝑃(𝑛) = Equation 12
𝑛!

Note that the probability of having no vehicles arrive in a time interval of length t, P(0), is
equivalent to the probability of a vehicle headway, h, being greater than or equal to the time
interval t. So from Eq. 12,
 −𝑞𝑡
𝑃(0) = 𝑃(ℎ ≥ 𝑡) = 𝑒 (3600) Equation 13

This distribution of vehicle headways is known as the negative exponential distribution and is
often simply referred to as the exponential distribution.

To help in visualizing the shape of the exponential distribution, Fig. 5 shows the probability
distribution implied by Eq. 13, with the flow, q, equal to 360 veh/h as in Example 6.

2.3.1 Limitations of Poisson’s Model

Empirical observations have shown that the assumption of Poisson-distributed traffic
arrivals is most realistic in lightly congested traffic conditions. As traffic flows become heavily
congested or when traffic signals cause cyclical traffic stream disturbances, other distributions of
traffic flow become more appropriate. The primary limitation of the Poisson model of vehicle
arrivals is the constraint imposed by the Poisson distribution that the mean of period
observations equals the variance. For example, the mean of period-observed traffic in Example
7 is 6.733 and the corresponding variance, σ2, is 7.210. Because these two values are close, the
Poisson model was appropriate for this example. If the variance is significantly greater than the
mean, the data are said to be overdispersed, and if the variance is significantly less than the
mean, the data are said to be underdispersed. In either case the Poisson distribution is no longer
appropriate, and another distribution should be used. Such distributions are discussed in detail
in more specialized sources [Transportation Research Board, 1975; Poch and Mannering, 1996;
Lord and Mannering, 2010].

Figure 5 Exponentially distributed probabilities of headways greater than or equal to t, with q = 360 veh/h.
2.4 SAMPLE COMPUTATIONS

Example 1 Flowrate
Let us suppose a 15-minute count of vehicles bound for Manila was conducted at a
particular location on Quezon Avenue. A summary is shown in the table below. Estimate the flow
rate in vehicles per hour.

TYPE 15 - MINUTE COUNT

Car/van 420
Jeepney 300
Bus 16
Truck 28

Solution:
The total number of vehicles counted in 15 minutes is 420 + 300 + 16 + 28 = 764. This is
expanded to one hour by multiplying by 4. Therefore, the flow rate is

q = 764 x 4 = 3,056 vehicles per hour

(Note that the count within the 15-minute interval is assumed to be the same for the other three
15-minute intervals.)

Example 2 Time -Mean Speed

The speed of 25 cars was observed. 10 cars were noted to travel at 35 kph, 8 cars at 40
kph, 2 cars at 50 kph, and 5 cars at 45 kph. Assuming that each car was traveling at constant
speed, determine the time mean speed.

Solution:
Applying equation 2:

∑𝑛𝑖=1(𝑢𝑖 ) (10𝑥35) + (8𝑥40) + (2𝑥50) + (5𝑥45)

𝑢̅𝑡 = = = 39.8 𝑘𝑝ℎ
𝑛 25

It is common practice among traffic engineers to report "spot speed" for given location.
Spot speed is often used as basis for establishing speed limits.
Example 3 Space -Mean Speed
The speed of 25 cars was observed. 10 cars were noted to travel at 35 kph, 8 cars at 40
kph, 2 cars at 50 kph, and 5 cars at 45 kph. Assuming that each car was traveling at constant
speed, determine the space mean speed.

Solution:
Applying equation 3:
𝑛 25
𝑢𝑖 = = = 39.26 𝑘𝑝ℎ
1 10 8 2 5
∑𝑛𝑖=1 ( ) + 40 + +
𝑢𝑖 35 50 45

Example 4 Headway
During morning peak hour, the average headway of UP-Katipunan jeepneys is estimated
at 5 minutes. If the passenger demand during the same period is 240, determine whether there
is a need to increase the number of jeepney units (or shorten the headway) for this route. Assume
that passenger demand is evenly distributed within that period and the average load/occupancy
is 14 passengers per jeepney. (Note: This assumption may not necessarily be true do to
fluctuation of passenger demand and variability of passenger occupancy.)
Solution:
From equation 7, the number of jeepneys per hour is
1
ℎ𝑡 = ;
𝑞
1 1
𝑞= = = 12 𝑗𝑒𝑒𝑝𝑛𝑒𝑦𝑠/ℎ𝑟
ℎ𝑡 5𝑚𝑖𝑛 (1 ℎ𝑜𝑢𝑟)
60𝑚𝑖𝑛

With an average load of 14 passengers per jeepney, the total number of passengers that can take
a ride is
12 jeepneys x 14 passengers/jeepney = 168 passengers

Since the demand during morning peak hour is 240 passengers, it can be said that there is a need
to increase the number of jeepney units during peak period.
Example 5 Spacing
During heavy traffic congestion, it was observed that the average spacing of vehicles in
queue in the innermost lane of EDSA is 6.5 m. Determine the jam density of stopped vehicles.

Solution:
1
𝑠= ;
𝑘
1 1 𝑣𝑒ℎ 154𝑣𝑒ℎ
𝑘= = = 153.85 ~
𝑠 6.5𝑚 (1 𝑘𝑚 ) 𝑘𝑚 𝑘𝑚
1000

Example 6 Vehicle Arrivals as a Poisson Process

An observer counts 360 veh/h at a specific highway location. Assuming that the arrival of
vehicles at this highway location is Poisson distributed, estimate the probabilities of having 0, 1,
2, 3, 4, and 5 or more vehicles arriving over a 20-second time interval.

Solution:
The average arrival rate, λ, is 360 veh/h, or 0.1 vehicles per second (veh/s). Using this in
Eq. 10 with t = 20 seconds, the probabilities of having exactly 0, 1, 2, 3, and 4 vehicles arrive are

(0.1𝑥20)0 𝑒 −0.1(20)
𝑃(0) = = 0.135
0!
(0.1𝑥20)1𝑒 −0.1(20)
𝑃(1) = = 0.271
1!
(0.1𝑥20)2 𝑒 −0.1(20)
𝑃(2) = = 0.271
2!
(0.1𝑥20)3 𝑒 −0.1(20)
𝑃(3) = = 0.180
3!
(0.1𝑥20)4 𝑒 −0.1(20)
𝑃(4) = = 0.090
4!

For five or more vehicles,

𝑃(𝑛 ≥ 5) = 1 − 𝑃(𝑛 < 5) = 1 − 0.135 − 0,271 − 0.271 − 0.180 − 0.090 = 0.053

A histogram of these probabilities is shown in figure 6.

 Figure 6 Histogram of the Poisson distribution for λ = 0.1 vehicles per second.

Example 7. Vehicle Arrivals as of Poisson Process with Detailed Vehical-Arrival Data

Traffic data are collected in 60-second intervals at a specific highway location as shown in
Table 1. Assuming the traffic arrivals are Poisson distributed and continue at the same rate as
that observed in the 15 time periods shown, what is the probability that six or more vehicles will
arrive in each of the next three 60-second time intervals (12:15 P.M. to 12:16 P.M., 12:16 P.M.
to 12:17 P.M., and 12:17 P.M. to 12:18 P.M.)?

Table 1 Observed Traffic Data

Solution:
Table 1 shows that a total of 101 vehicles arrive in the 15-minute period from 12:00 P.M.
to 12:15 P.M. Thus, the average arrival rate, λ, is 0.112 veh/s (101/900). As in Example 6, Eq. 10
is applied to find the probabilities of exactly 0, 1, 2, 3, 4, and 5 vehicles arriving. Applying Eq. 10,
with λ = 0.112 veh/s and t = 60 seconds, the probabilities of having 0, 1, 2, 3, 4, and 5 vehicles
arriving in a 60-second time interval are (using λt = 6.733).

(6.733)0 𝑒 −(6.733)
𝑃(0) = = 0.0012
0!
(6.733)1𝑒 −(6.733)
𝑃(1) = = 0.008
1!
(6.733)2 𝑒 −(6.733)
𝑃(2) = = 0.027
2!
(6.733)3 𝑒 −(6.733)
𝑃(3) = = 0.0606
3!
(6.733)4 𝑒 −(6.733)
𝑃(4) = = 0.102
4!
(6.733)5 𝑒 −(6.733)
𝑃(5) = = 0.137
5!

The summation of these probabilities is the probability that 0 to 5 vehicles will arrive in any given
60-second time interval, which is

𝑃(𝑛 ≤ 5) = ∑ 𝑃(𝑛)
𝑖=0
𝑃(𝑛 ≤ 5) = 0.0012 + 0.008 + 0.027 + 0.0606 + 0.102 + 0.137 = 0.3358
So 1 minus 𝑃(𝑛 ≤ 5) = is the probability that 6 or more vehicles will arrive in any 60-second
interval, which is

𝑃(𝑛 ≤ 5) = 1 − 𝑃(𝑛 ≤ 5)
𝑃(𝑛 ≤ 5) = 1 − 0.3358 = 0.6642

The probability that six or more vehicles will arrive in three successive time intervals (t1,t2, and
t3) is simply the product of the probabilities, which is
3

𝑃(𝑛 ≥ 6)𝑓𝑜𝑟 3 𝑠𝑢𝑐𝑐𝑒𝑠𝑠𝑖𝑣𝑒 𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑𝑠 = ∏ 𝑃(𝑛 ≥ 6)

𝑡𝑖 =1
𝑃(𝑛 ≥ 6)𝑓𝑜𝑟 3 𝑠𝑢𝑐𝑐𝑒𝑠𝑠𝑖𝑣𝑒 𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑𝑠 = (0.6642)3 = 0.293
Example 8. Headways and the Negative Exponential Distribution
Consider the traffic situation in Example 6 (360 veh/h). Again assume that the vehicle
arrivals are Poisson distributed. What is the probability that the headway between successive
vehicles will be less than 8 seconds, and what is the probability that the headway between
successive vehicles will be between 8 and 10 seconds?

Solution:

By definition, 𝑃(ℎ < 𝑡)= 1-𝑃(ℎ ≥ 𝑡). This expression gives the probability that the
headway will be less than 8 seconds as

−𝑞𝑡
𝑃(ℎ < 𝑡) = 1 − 𝑒 (3600)

−360(8)
( )
𝑃(ℎ < 𝑡) = 1 − 𝑒 3600 = 0.551

To determine the probability that the headway will be between 8 and 10 seconds, compute the
probability that the headway will be greater than or equal to 10 seconds:

−𝑞𝑡 −360(10)
(
𝑃(ℎ ≥ 𝑡) = 𝑒 (3600)
=𝑒 3600 ) = 0.368

So the probability that the headway will be between 8 and 10 seconds is 0.081 (1 − 0.551 − 0.368).

References
Garber, N. J., & Hoel, L. A. (2019). Traffic and highway engineering. Cengage Learning
Homburger, W. S., Hall, J. W., Reilly, W. R., & Sullivan, E. C. (2007). Fundamentals of traffic engineering.