Autonomous vehicles (AVs) have the potential to improve automotive safety, largely by removing human error. However, achieving lane-keeping control for AVs is a task crucial to safety that is susceptible to intermittent malfunctions of vision sensors due to lighting changes, incomplete lane marks, shadows, etc. Different from robust lane detection algorithms, this paper treats these abnormal conditions as intermittent sensor faults and aims to present a robust lane-keeping control scheme from the viewpoint of the fault detection (FD) and fault-tolerant control (FTC). The core idea behind this method is the integrated design of the observer gains and the feedback gains that incorporate the fault duration and convex design of the controller parameters by solving a set of linear matrix inequalities (LMIs). The proposed control scheme guarantees that the closed-loop system maintains input-to-state stability (ISS), while the error signals theoretically converge to a small neighborhood of the origin. The effectiveness of the proposed approach is confirmed by experimental results for real vehicles.
Fault-Tolerant Lane-Keeping Control of Autonomous Vehicles Subject to Intermittent Malfunctions of Vision Sensors
2023-10-27
2716888 byte
Conference paper
Electronic Resource
English
Accommodating Controller Malfunctions Through Fault Tolerant Control Architecture
Online Contents | 2007
|