Lane Following for Autonomous Vehicles - Novel Heading Error Definition and Evaluation of Discrete vs. Continuous Time Control

Abstract: The autonomous vehicle has a multilayer controlarchitecture for determining its movements, in the top layer aplanner reads the environment and decides on a path and speedat which to drive, it is then up to a quicker and less complexcontroller to make sure that the generated lane is followed.The thesis considers a controller system for lane following ofan autonomous vehicle. The first purpose of the thesis is toanalyze the performance of a lane following controller, usinga new heading error definition. The second is to investigate thedifference in performance between a controller that accounts for,and one that does not account for, the sample rate of the discretetime system representing the actual sampled digital controller. Itis shown that the new heading error definition gives reliableresults in both normal and extreme situations and that thedifference in performance when considering the sample rate ornot is very small.