Stability Analysis of a Longitudinal Control Law for Autonomous Vehicles
Abstract
We focus in this article on the analysis of an existing acceleration law for a following autonomous vehicle with constant time headway goal. We first define a function that measures the positioning of the following vehicle compared to the leading vehicle with respect to the desired time headway. We study the convergence and bounds of this positioning function and prove some key results in an ideal case, i.e., when all accelerations are unbounded. Then we conduct several simulations that validate our theoretical results and show their limitations in real cases, i.e., when the accelerations are technologically bounded. Finally we use decoupling and feedback linearization techniques to derive a new acceleration law whose convergence is always exponential in the ideal case, and we compare both laws.