The well known road detection and tracking algorithm (RDT), developed at the Universitat der Bundeswehy Muchen (UBM), has been adapted for following unpaved paths and contour lines. The vision system consists of a color CCD-camera mounted on UBM's high bandwidth pan-tilt head, dubbed TACC. The monochrome and color signals from the camera are processed in parallel. This vision system is used for lateral control of the Primus-C experimental vehicle Digitized Wiesel 2, an air-transportable tracked tank. Image processing simultaneously exploits the results from edge-based feature extraction and area-based segmentation. Feature matching is facilitated by exploiting the photometric homogeneity along contours. A flexible algorithm for active viewing direction control was implemented. The two axis camera carrier (TACC) is controlled by fixating on a point moving on the ceterline of the road segment. A semi-autonomous initialization mode is integrated using image points on the contour to be followed, specific by an operator. Autonomous driving speeds up to 50km/h on unpaved roads was demonstrated publicly in June 1999.
[1]
Stefan Baten.
Autonomous road and contour following with a tracked vehicle
,
2000,
Defense, Security, and Sensing.
[2]
Markus Maurer,et al.
A LOW COST VISION SYSTEM FOR AUTOMOTIVE APPLICATIONS
,
1997
.
[3]
Reinhold Behringer.
Real-time 3D reconstruction of road curvature in far look-ahead distance from analysis of image sequences
,
1995,
Other Conferences.
[4]
Ingo Schwartz.
PRIMUS: realization aspects of an autonomous unmanned robot
,
1998,
Defense, Security, and Sensing.
[5]
E. D. Dickmanns,et al.
Scene Recognition and Landmark Navigation for Road Vehicles
,
1995
.
[6]
Michael Lützeler,et al.
ROAD RECOGNITION WITH MARVEYE
,
1998
.
[7]
Ernst D. Dickmanns,et al.
An Expectation-based, Multi-focal, Saccadic (EMS) Vision System for Vehicle Guidance
,
2000
.