TY - JOUR
T1 - A New Approach to Determine Traffic Peak Periods to Utilize in Transportation Planning
AU - Saracoglu, Abdulsamet
AU - Ozen, Halit
AU - Apaydin, Mehmet Serkan
AU - Maltas, Abdullah
N1 - Publisher Copyright:
© 2021, King Fahd University of Petroleum & Minerals.
PY - 2021/11
Y1 - 2021/11
N2 - Detection of the traffic peak periods for a city has many advantages not only in terms of the traffic operation but also in urban transportation planning. In the planning, future travel demand is analyzed for peak and off-peak periods separately and transportation networks are designed based on these analyses. However, the traffic peak periods may shift in time because of changes in people's activities. Therefore, the detection and update of these periods periodically provide to improve urban transportation planning. The aim of this research is to present a new approach to determine traffic peak periods in order to utilize for use in transportation planning. In order to achieve this goal, the inductive loop detector data are obtained for a city and cleaned. Regular traffic flow patterns (without days that include unexpected incidents) of the whole city are produced, and peak periods are determined with these data. The Facebook Prophet software, which is a forecasting procedure, is presented as a new approach to determine traffic peak periods based on changepoint detection. In order to compare the results, experts' opinion, which is the conventional method, and the k-medoids clustering methods are applied. In conclusion, it is seen that the suggested new approach facilitates so much the accurate determination of traffic peak periods for urban transportation planning.
AB - Detection of the traffic peak periods for a city has many advantages not only in terms of the traffic operation but also in urban transportation planning. In the planning, future travel demand is analyzed for peak and off-peak periods separately and transportation networks are designed based on these analyses. However, the traffic peak periods may shift in time because of changes in people's activities. Therefore, the detection and update of these periods periodically provide to improve urban transportation planning. The aim of this research is to present a new approach to determine traffic peak periods in order to utilize for use in transportation planning. In order to achieve this goal, the inductive loop detector data are obtained for a city and cleaned. Regular traffic flow patterns (without days that include unexpected incidents) of the whole city are produced, and peak periods are determined with these data. The Facebook Prophet software, which is a forecasting procedure, is presented as a new approach to determine traffic peak periods based on changepoint detection. In order to compare the results, experts' opinion, which is the conventional method, and the k-medoids clustering methods are applied. In conclusion, it is seen that the suggested new approach facilitates so much the accurate determination of traffic peak periods for urban transportation planning.
KW - Changepoint detection
KW - Facebook prophet
KW - k-medoids clustering
KW - Traffic peak period
KW - Transportation planning
UR - http://www.scopus.com/inward/record.url?scp=85100430171&partnerID=8YFLogxK
U2 - 10.1007/s13369-021-05384-2
DO - 10.1007/s13369-021-05384-2
M3 - Article
AN - SCOPUS:85100430171
SN - 2193-567X
VL - 46
SP - 10409
EP - 10418
JO - Arabian Journal for Science and Engineering
JF - Arabian Journal for Science and Engineering
IS - 11
ER -