Modeling Lateral and Longitudinal Control of Human Drivers with Multiple Linear Regression Models

Abstract

In this paper, we describe results to model lateral and longitudinal control behavior of drivers with simple linear multiple regression models. This approach fits into the Bayesian Programming (BP) approach (Bessière, 2008) because the linear multiple regression model suggests an action selection strategy which is an alternative to the BP action selection strategies draw and best. Furthermore, the inference process provided by a linear multiple regression model is a kind of short cut inference compared to the inference approach used in Bayesian networks or Bayesian Programming.