Safety of the Stanley controller with curved lanes and noisy perception: Theories and simulations

The Stanley controller was designed for the DARPA 2005 Grand Challenge and has been widely used in real autonomous vehicles and simulation models. While the original paper presented an analysis of the tracking performance of this controller in straight roads, the analysis for general curved lanes and with perception errors is not available. This thesis utilizes the Lyapunov theory and the entropy estimation theory to give tracking performance guarantees for the Stanley controller. The analysis in this thesis can be used for ensuring safety of lane following and provide guidelines for choosing design parameters with respect to velocity, lane curvation, and sensing constraints. Furthermore, the original Stanley controller just assumes a constant velocity for the vehicle and thus doesn’t contain any form of longitudinal control. The analysis in this thesis also contains theoretical work to create an optimal velocity profile which can enable the car to achieve smaller deviations from the reference path and shorter track completion time. In addition, this thesis has simulations for its theoretical work.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Electrical & Computer Engr

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2023

Author dc:creator

• Li, Hongyi

Contributors dc:contributor

• Mitra, Sayan

• Control Theory
• Autonomous Driving

Statement dc:rights

• Copyright 2023 Hongyi Li

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/122080

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F122080