This paper proposes a new path tracking control system that incorporates an image processing system for detecting forward lane markers by means of a vehicle-mounted camera. The new tracking method is proposed from the standpoint of modern control theory: The new image processing system proposed here incorporates a method for considering the host vehicle's dynamic characteristics into an extended Kalman filter, which is used for estimating the road curvature and state quantities of the vehicle. By using steering angle information, this method makes it possible to provide both good convergence of the estimated values and improved anti-noise performance. The attainment of both capabilities was confirmed by theoretical calculations and the results of actual driving tets. An optimum steering control system is also proposed which is constructed such that the image processing system directly estimates the dynamic characteristics of the host vehicle and uses the results for steering control. The effectiveness of this system was verified by computer simulations.
Investigation of automatic path tracking using an extended Kalman filter
Untersuchung der automatischen Spurverfolgung mit einem erweiterten Kalman-Filter
JSAE Review ; 23 , 1 ; 61-67
2002
7 Seiten, 10 Bilder, 3 Quellen
Aufsatz (Zeitschrift)
Englisch
Investigation of automatic path tracking using an extended Kalman filter
Online Contents | 2002
|Investigation of automatic path tracking using an extended Kalman filter
Kraftfahrwesen | 2002
|Maneuvering target tracking using extended Kalman filter
IEEE | 1991
|