Review of Model Predictive Control and Other Methodologies for Lateral Control of an Autonomous Vehicle

Abstract

In this chapter, we present a study on autonomous vehicle path tracking utilizing both Linear, Linear Time Varying, Nonlinear Model Predictive controllers and evaluate their performance on a given path. The vehicle is both a linear and non-linear bicycle model with front steering . The performance metric in the evaluation of the controller/path tracker is the root mean square error between the defined path and the resulting vehicle path. The path tracking algorithm utilized are Serret-Frenet Equations and a modified version of Pure Pursuit. The robustness of the controllers are examined by varying vehicle mass, tire-road coefficient of friction, and tire cornering stiffness.