AMUSE

Date: October 10, 2018

Time: 6:00PM - 7:00PM

Location: BLOC 220

Speaker: Dr. Reza Langari, Department of Mechanical Engineering, Texas A&M University

Title: Computational Model of Human Driving for Autonomous Driving

Abstract: In this presentation, we consider how a combination of mathematical tools ranging from predictive modeling to game theory to Markov decision processes can be used to devise suitable models of human driving for automated vehicles. The intent is to show how human driving can be modeled in algorithmic form and used to design collision free path planning and control strategies for these vehicles. The use of mathematical tools is essential in this process as one must have efficient ways to represent human driving behavior. For instance, a multitude of rules of thumb will not be sufficient for the purpose at hand. These algorithms are matched against human driver behavior in driving simulator studies, thereby proving their effectiveness. We demonstrate the utility of the approach using high fidelity simulation models of multilane highway driving in mandatory and discretionary lane changing settings.