Doctor

Framework for estimating congestion performance measures: from data collection to reliability analysis. Case study Stockholm

Carlos A. Morán Toledo

Licentiate Thesis

Division of Traffic and Logistics Department of Transport and Economics School of Architecture and the Built Environment Royal Institute of Technology

Stockholm, Sweden

ABSTRACT
For operational and planningpurposes it is important to observe and predict the traffic performance of congested urban road links and networks. Congestion can be defined as traffic conditions caused by a downstream bottleneck or excess in travel time from what is incurred during light or free-flow travel conditions. Factors affecting definitions of congestion for specific studies are reviewed and an inventory of proposedcongestion performance measures is presented for both definitions of congestion. Swedish Road Administration has recognized the reliability of the estimations of congestion levels as an important factor when describing the traffic condition in the road traffic network. Traffic data collected for the Stockholm congestion charging trials was used to estimate selected congestion performance measuresand to analyze their statistical characteristics and applicability. A comparative analysis of data collection methods is provided and further recommendations for their using estimating Congestion Performance Measures. The reliability of the estimations of each Congestion Performance Measure is evaluated for different area networks and different time periods of the day. These series ofobservations are further studied aiming to identify systematic differences in the reliability of the estimations. A reliability ranking is provided for guiding future studies in the selection of estimators. Further, a simplify methodology for estimating recommended sample sized under budget restricted conditions is provided.

ACKNOWLEDGMENTS
I would like to express my sincere gratitude to mysupervisor, Professor Karl-Lennart, for his encouragement patience and excellence guidance. Karl-Lennart’s enthusiasm and devotion towards traffic engineering have been of inestimable importance for the completion of this work. I would like to thanks Hans Cedermark from CDU that supported me in the early stages of this project. My colleagues at the department of traffic and logistic are acknowledge forcreating a stimulating work place. I would like also to specially thanks Karin Brundell-Frej for providing me with her positive discussions and rich comments This work has been found by the Swedish road administration through the project “trängselmått I” and trängselmått II”. This financial support is gracefully acknowledged. Finally I would like to thank my family who have supported medthrough my stay in Sweden, my friends for supporting me and making my integration in Sweden easier.

Stockholm, January 2008

Carlos Morán T.

TABLE OF CONTENTS
ABSTRACT ACKNOWLEDGEMENTS I PROBLEM DESCRIPTION ......................................................................1 I.1 I.2 I.3 II BACKGROUND......................................................................................1 OBJECTIVES ........................................................................................1 SCOPE ................................................................................................2

LITERATURE STUDY ..............................................................................3 II.1 THEORETICAL FRAMEWORK FOR STATISTICAL ANALYSIS ...........................3 II.2MODELLING THE EFFECTS OF CONGESTION.............................................4 II.2.1 Deterministic approach...............................................................5 II.2.2 Stochastic approach ...................................................................6 II.2.3 Simulation models ......................................................................6 II.3 DEFINITION OF CONGESTION...