Date of Award

8-2013

Document Type

Thesis

Degree Name

Master of Science (MS)

Legacy Department

Civil Engineering

Committee Chair/Advisor

Chowdhury, Mashrur

Committee Member

Khan , Taufiqaur

Committee Member

Ogle , Jennifer

Abstract

Both safety and the capacity of the roadway system are highly dependent on the car-following characteristics of drivers. Car-following theory describes the driver behavior of vehicles following other vehicles in a traffic stream. In the last few decades, many car-following models have been developed; however, studies are still needed to improve their accuracy and reliability.
Car-following models are a vital component of traffic simulation tools that attempt to mimic driver behavior in the real world. Microscopic traffic simulators, particularly car-following models, have been extensively used in current traffic engineering studies and safety research. These models are a vital component of traffic simulation tools that attempt to mimic real-world driver behaviors. The accuracy and reliability of microscopic traffic simulation models are greatly dependent on the calibration of car-following models, which requires a large amount of real world vehicle trajectory data.
In this study, the author developed a process to apply a stochastic calibration method with appropriate regularization to estimate the distribution of parameters for car-following models. The calibration method is based on the Markov Chain Monte Carlo (MCMC) simulation using the Bayesian estimation theory that has been recently investigated for use in inverse problems. This dissertation research includes a case study, which is based on the Linear (Helly) model with a different number of vehicle trajectories in a highway network.
The stochastic approach facilitated the calibration of car-following models more realistically than the deterministic method, as the deterministic algorithm can easily get stuck at a local minimum. This study also demonstrates that the calibrated model yields smaller errors with large sample sizes. Furthermore, the results from the Linear model validation effort suggest that the performance of the calibration method is dependent upon size of the vehicle trajectory.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.