Research
Background:
It
has long been known that platoons, which form
at the exit of a given traffic signal, do not
remain intact or compact as they progress
along an arterial link towards the next
traffic signal.
Platoons may disperse along the road
either more quickly or slowly depending on the
actual road geometric and traffic conditions
between the two adjacent intersections of
interest. In part, this dispersion of vehicle
platoons occurs due to the differences in the
desired speeds of the various drivers that
make up the platoon. However, a large portion
of the dispersion is also caused by the fact
that some vehicles will experience delays
while travelling along the link which are
random in terms of both their occurrence and
duration.
The majority of these random delays
along links occur when vehicles slow down for
other vehicles, which are either turning of
the road at a mid-block location, or
attempting to enter or leave on-street
parking.
The
calculation of delays and stops of coordinated
traffic signals by both off-line and on-line
models such as the widely used TRANSYT
(Robertson, 1969) and SCOOT (Hunt, et al.,
1981) relies on the model’s ability to
accurately predict traffic flow patterns from
one signal to another. The effectiveness of
the coordinated signal timings depends on the
accuracy of the calculated delays and stops
and thus on platoon dispersion modeling. At
present, one of the most commonly utilized
macroscopic approaches to the modeling of the
platoon dispersion process is the one
developed by Robertson (1969), which was later
incorporated into the TRANSYT.
This approach has since also become a
virtually universal standard throughout the
world in other control or simulation models
such as SCOOT, SATURN (Hall, et al., 1980),
TRAFLO (Lieberman, et al., 1980), and
INTEGRATION (Van Aerde and Yagar, 1990).
A
successful application of the Robertson's
platoon dispersion model to modeling platoon
dispersion relies on the appropriate
calibration of several model parameters. The
empirical studies performed by the Transport
and Road Research Laboratory (TRRL) in the
United Kingdom suggested some default values
for the platoon dispersion modeling. The work
performed by PRC Engineering (Tarnoff and
Parsonson, 1981) and the University of Florida
(Lorick, et al., 1980) suggested a set of
default values for the platoon dispersion
parameters for the North American version of
TRANSYT-7F. Although many research findings
have indicated that the platoon dispersion
parameters should be site-specific and a
function of the road grades, curvature,
parking, opposing flow interference, traffic
volume and other sources of impedance, no
methodology exists that can quantitatively
calibrate the platoon dispersion parameters.
The
continuing use of the default values for the
platoon dispersion modeling may risk the
implementation of virtually ineffective signal
timing plans on roads. With the growing
applications and development of the Advanced
Traffic Management Systems (ATMS) throughout
the world, more and more real-time control
systems for coordinated traffic signals are
expected to be deployed on various urban
networks. Therefore, it becomes even more
critical for traffic engineers to use the
accurate modeling approach for determining the
coordinated signal timing plans in order not
to waste the resources for investment and
deployment of the advanced traffic control and
management systems. The calibration of the
platoon dispersion parameters is one of the
central issues affecting the efficacy of the
coordinated signal timings.
Research
Objectives
The
objectives of this research are threefold.
In the first instance, the research
will examine in a great detail the underlying
assumptions of the TRANSYT macroscopic platoon
dispersion model. The examination will bring
into question the common assumption that b
equals 0.8 for all values of a.
In the second instance, the research
will develop an alternate mathematical
approach for calibrating the platoon
dispersion parameters directly from the
statistical properties of the travel time
experiences of individual vehicles, which can
be obtained on a real-time basis in ATMS
applications. Finally, the research will
collect link travel time data from selected
roads and calibrate the platoon dispersion
parameters using the proposed approach, thus
establishing the context for using the
proposed calibration approach in real world
applications.
Research
Abstract
Vehicles
form platoons at the exit point of a given
traffic signal, which will disperse while they
progress along the link towards the next
downstream traffic signal. The platoons may
disperse either more quickly or slowly
depending on the actual road geometric and
traffic conditions between the two adjacent
intersections of interest. The adequate
modeling and description of the platoon
dispersion behavior ultimately affect the
quality of the coordinated traffic signal
timings.
At present, the most widely used
modeling method of platoon dispersion is the
TRANSYT’s macroscopic platoon dispersion
model in which the determination of its major
parameters is based on the empirical values.
This report presents a methodology for
calibrating the platoon dispersion parameters
in the TRANSYT’s platoon dispersion model,
which is based on a statistical analysis of
link travel time data rather than more
traditional goodness-of-fit tests between the
observed and the projected vehicles’
progression patterns.
Specifically, the platoon dispersion
parameters are made explicit dependent
variables of the average link travel time and
the standard deviation of link travel times.
The proposed technique is suited for
applications in advanced traffic management
systems (ATMS) networks where the required
link travel time data could be obtained on a
real-time basis.
The calibration of platoon dispersion
parameters using the proposed technique for
the field collected data has shown that
platoon dispersion parameters are indeed
different, even on the same street but with
different travel times.
This conclusion confirms the need of
calibrating platoon dispersion parameters on a
link specific basis.
Research
Report
The
research report has been published by
Southwest Region University Transportation
Center (SWUTC) at Texas Transportation
Institute (TII) with a number SWUTC/99/472840-00044-1.
The research was also presented at TRB 2000
Annual Meeting (00-0841) and will appear in
the forthcoming Transportation Research
Record. To order a copy of report,
please contact Dr. Lei Yu by telephone at
(713) 313-7282 or by e-mail at
yu_lx@tsu.edu.
