Title :
Estimation of Lane Marker Parameters With High Correlation to Steering Signal
Author :
Demcenko, A. ; Tamosiunaite, Minijia ; Vidugiriene, Ausra ; Jakevicius, L.
Author_Institution :
Dept. of Appl. Inf., Vytautas Magnus Univ., Kaunas, Lithuania
fDate :
6/1/2012 12:00:00 AM
Abstract :
This paper considers the design and analysis of lane marker parameters in 2-D images. Five different parameters, which have high correlation to the steering angle of a vehicle, are proposed, and their correspondence to the steering signal is analyzed. The parameters are based on the position, angle, area, and curvature of the lane marker and have not been reported in the literature before, except the curvature. A new derivative-free method is proposed for curvature estimation. The stability of the proposed parameters is analyzed with respect to the look-ahead distance. Possible application of the parameters for overtaking detection is presented. This paper is performed using signals from real country road driving.
Keywords :
driver information systems; object detection; parameter estimation; stability; steering systems; 2D images; curvature estimation; derivative-free method; lane marker parameter analysis; lane marker parameter estimation; overtaking detection; parameter stability; real country road driving; steering signal analysis; vehicle steering angle; Correlation; Estimation; Roads; Solids; Stability analysis; Vectors; Vehicles; Driver´s assistance; human-like driving; lane marker; steering signal;
Journal_Title :
Intelligent Transportation Systems, IEEE Transactions on
DOI :
10.1109/TITS.2011.2179535