BSc Final Project
Traffic state estimation with multi-sensor data for large networks with macro simulation
Yazan Safadi
The project was divided into three parts: the first and second were supervised by Prof. Nikolas Geroliminis, during Yazan's exchange year at EPFL (Lausanne); the third was supervised by Prof. Jack Haddad at Technion. The project was recognized as excellent and was awarded The "Eng. Yitzhak Hershkovitz" prize.
PART A - This project intends to identify which should be the optimal distribution of traffic sensors to estimate the Macroscopic Fundamental Diagrams (MFD) using sample data from the city of Barcelona, finding differences and comparing the data obtained for different percentages and distributions of sensors in the studied network. This study has tested how the number of detectors affects the MFD behaviour and the critical point in which the network reaches congestion. (The thesis and presentation slides files for part A are available below)
PART B - In this project, we investigate the MFD model for a Perimeter Control approach in a microsimulation environment (Aimsun software). Model estimation was compared with the microsimulation results when estimating the traffic state. (The presentation slides file for part B is available below)
PART C - In this project, we explored the MFD model in a microsimulation environment (Aimsun software), and looked at different networks to observe how they affected the MFD curve. The MFD variables were extracted by developing API code utilizing the Aimsun API. A MFD curve was produced by analyzing the data later in MATLAB. The developed API code was an initial step toward developing a Perimeter Control approach and testing it in a microsimulation environment. (The presentation slides file for part C is available below)
Abstract
Mobility is increasing and traffic management is becoming a key factor to successfully manage any urbanized region. In relation to that, a good knowledge of the state of the network is essential for traffic regulators, policy makers and final users. This project intents to identify which should be the optimal distribution of traffic sensors using sample data from the city of Barcelona, finding differences and comparing the data obtained for different percentages and distributions of sensors in the studied network. Finally, a methodology to identify the best detectors is recommended, and by analysis data from Aimsun it is possible to measure the number of detectors in the network unrelated to the detectors locations, which will represent and control the whole network