Share Email Print

Proceedings Paper

Determination of an autonomous vehicle's position by matching the road curvature
Author(s): Chao-Chi Huang; Duncan Fyfe Gillies
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

This paper presents a solution to the problem of position estimation for autonomous land vehicles (ALV). In our previous research, we have used the recursive least square error minimization method to determine the vehicle's pose information. When a reasonably precise initial estimate is used, our system will converge to the vehicle's pose parameters with high accuracy. Using the global positioning system (GPS), we can obtain a position estimate with errors which, in normal conditions, are of the order of 100 meters. This accuracy is not sufficient for automated navigation. The initial estimate taken from the GPS system can be refined and improved by using curvature matching. Comparing the expected and calculated 3D road curvature, the system can recognize the current position on the road. Curvature in the 3D space is determined by selecting and backprojecting points from a 2D road image and fitting a set of cubic spline patches to them. At present the system has been tested using both synthetic data and real data. The initial results indicate that the method can be made to work well, however, care is required in the measurement and calculation of the curvature in real applications.

Paper Details

Date Published: 27 December 1995
PDF: 11 pages
Proc. SPIE 2591, Mobile Robots X, (27 December 1995); doi: 10.1117/12.228982
Show Author Affiliations
Chao-Chi Huang, Imperial College of Science, Technology and Medicine (United Kingdom)
Duncan Fyfe Gillies, Imperial College of Science, Technology and Medicine (United Kingdom)

Published in SPIE Proceedings Vol. 2591:
Mobile Robots X
William J. Wolfe; Chase H. Kenyon, Editor(s)

© SPIE. Terms of Use
Back to Top