TITLE:

Human/Machine Learning in Driver Behavior Modeling

DATE:

Friday, February 12th, 2016

TIME:

3:30 PM

LOCATION:

GMCS 214

SPEAKER:

Sahar Ghanipoor Machiani

ABSTRACT:

Traffic conflicts associated to signalized intersections are one of the major contributing factors to crash occurrences. Driver

behavior plays an important role in traffic safety at signalized intersections. In this seminar, dynamics of driver behavior in relation to the

traffic conflicts occurring at the onset of yellow is discussed. The area ahead of intersections in which drivers encounter a dilemma to pass through

or stop at the onset of yellow indication is called dilemma zone (DZ). The focus of my research is on drivers’ decision dynamics, human learning, and

choice behavior in DZ, and DZ-related safety measures. I developed an adaptive experimental design in a driving simulator to capture drivers learning

process while experiencing safe and unsafe signal settings. I developed an agent-based human learning model integrating machine learning and human

learning techniques. An abstracted model of human memory and cognitive structure was used to model agent’s behavior and learning. I will discuss the

results of the model which was trained using the driving simulator data. Next, the possibility of predicting drivers’ decision approaching a yellow

signal indication in different time frames will be discussed. I applied a machine learning technique -discriminant analysis- to construct a driver

behavior prediction model. This model assists advanced signal protection algorithms to make more intelligent decisions.

HOST:

Dr. Jose Castillo

DOWNLOAD: