Predictive algorithms for train arrival and clear times at highway-rail intersections /

Research is being performed to integrate railroad information

Bibliographic Details
Main Author: Halloin, David Muyskens
Format: Thesis Book
Language:English
Published: [Place of publication not identified] : [publisher not identified] ; 1997.
Subjects:
Description
Summary:Research is being performed to integrate railroad information
into traffic management centers for the improvement of
traffic operations and safety on the surface street system.
An issue of major concern is the advanced warning of train
arrivals at highway-rail intersections. Several research
projects are being performed to develop new technologies for
gathering this information from the rail corridor and for
communicating it to traffic management centers. Little or no
research has been done to incorporate existing technologies
for this purpose. Lack of methods for using technologies
already available in the field to predict train arrivals has
prompted this research focusing on the development of
algorithms, using data obtainable from existing technologies,
for prediction of train arrival and clear times. A typical
rural-to-urban rail corridor, containing numerous at-grade
highway-rail intersections, was chosen for this study. Six
general algorithms, incorporating different types and amounts
of information, were developed to predict train arrival
times. Four algorithms for predicting intersection blockage
by trains were developed to be used in predicting clear
times. These algorithms were evaluated based on their logic,
ability to explain actual train movements, as well as their
accuracy and precision. Results indicated that predictive
algorithms incorporating the largest amount of information
yielded the most accurate results. It was found that
acceleration and deceleration rates of trains between the
rural detection intersections included in the study were not
useful in predicting arrival times downstream for the study
corridor. Train speeds at the detection intersections,
however, are useful in making these predictions. Comparisons
between the use of actual and mean historic train speeds at
detection intersections indicated a distinct advantage in the
use of actual train speeds. Comparisons between using link-
speed and point-speed detection methods showed an advantage
in precision and accuracy with the use of the point-speed
method. For the study corridor, it was found that the use of
posted train speeds in algorithms consistently led to better
predictions than did the use of mean historical train speeds.
Item Description:Vita.
"Major subject: Civil Engineering".
Physical Description:xi, 104 leaves : illustrations ; 28 cm.
Issued also on microfiche from Lange Micrographics.
Bibliography:Includes bibliographical references.