Abstract:
The knowledge of travel times on road networks has vital importance to measure
network performance and quantify congestion for road users as well as for network
operators. Congestion of an urban road network varies in different times of a
weekday and weekends. Time varying congestion heat maps are easily understood
by all transport users. And time varying contour maps are effective way to estimate
the network performance at various times by transportation planners. In this study
variation of travel time on a selected road network of Dhaka city is estimated by
developing congestion heat maps and contour maps by using MATLABprogramming
software. And also shortest possible route for a certain Origin -
Destination pair has been tested by using shortest path algorithm. Average
Intersection delay is estimated for both week days and weekend days in this study.
In this study, travel time data collection technique has been selected, the data
collection by GPS-enabled devices such as, smart phone. GPS enabled vehicles
provide an alternative to other travel time data collection methods with decent
accuracy and can be used for providing real time traffic condition of the network. As
Dhaka city has the high penetration of smart phone market so it is possible to extract
a huge data set of travel time of a road network by this data collection method.
The data processing and data analysis part is mainly done by the programming
software- MATLAB which processes huge set of data in a short time integrating
with visualization.
Generated heat maps and contour maps shows a significant change in congestion
from 1st slot (8.00.00AM- 11.59.59AM) to 2nd slot(12.00.00PM-3.59.59PM) , 3rd
(4.00.00 PM-7.59.59PM) and 4th slot (8.00.00PM-10.59.59PM) on weekday. And
for weekend maps shows congestion criteria is mostly haphazard pattern, though
most of the congestion occurs on 2nd slot (12.00.00PM-3.59.59PM) and 4th slot
(8.00.00PM-10.59.59PM) on weekend.
The shortest path estimation for the O-D pair 48(23.742651, 90.395722, Ruposhi
Bangla Mor) to 37(23.727629, 90.410457, Shohid Noor Hossain Square) is done by Dijkstra's Algorithm, which shows a shortest path for this pair varies in weekday but
is almost same for weekend on this pair.
As this study shows the route choice of the road users can vary in different time if
the real time congestion maps are available. Thus this model developed in this thesis
can be used in further study to develop a real time congestion analysis for the full
network of any congested city in Bangladesh and construct route choice tools for the
road users.
