A.3.2 Skidpan demo to Daimler-Benz AG in 1986
The first six years of developing vision for road vehicles had been done at UniBwM without any partner from outside; after the principle solution had been found in simulations, demonstrations in the real world were envisaged. In 1985 a 5-ton van DB 508 dubbed ‘VaMoRs’ has been equipped as test vehicle for computer vision by the “Institut fuer Systemdynamik und Flugmechanik”, LRT; a second version of the 1st -generation image processing system BVV2, custom designed by the “Institut fuer Messtechnik” (Prof. V. Graefe), and realized with 16-bit micro-processor technology (Intel 80x86) together with a PC for system control provided the data processing capabilities for analyzing several subregions (windows) of every second video field (320x120 pixels, at 12.5 Hz).
H.2.3 Van VaMoRs (1985 – 2004, three phases)
With these systems in place at UniBwM, the research division of Daimler-Benz AG became interested in cooperating in the field of future “ground vehicles capable of vision”. Therefore, a demonstration of the state of the art developed at UniBwM was organized in December 1986.
Both longitudinal and lateral control based on visual recognition of the environment was to be shown. VaMoRs was standing ‘ready to go’ at the entrance of the skidpan when the system was switched into the autonomous mode; along a straight line painted on the ground, the vehicle had to accelerate to about 10 m/s (36 km/h), see top left of figure. From the end of this straight section a spiral curve should have been painted as guide line leading tangentially to the border line between a brighter circular region outside (concrete) and a darker region inside (dark blue basalt, see screenshots below). Unfortunately, the transition onto the circle realized showed a slight kink (finite negative angle, see figure); the decision was taken to stick with this imperfectness and to see what the vision system would do. Results are described in ([Dickmanns 2007], Section 7.3.3, p.214); an impression of the performance achieved may be obtained from video 02.
video 02 SkidpanDB-VaMoRs 1986
Dickmanns ED, Zapp A (1987). Autonomous High Speed Road Vehicle Guidance by Computer Vision. 10th IFAC World Congress Munich, Preprint Vol. 4, pp 232–237
Graefe V (1984). Two Multi-Processor Systems for Low Level Real-Time Vision. In Brady et al. (eds. 1984). Robotics and Artificial Intelligence, Springer-Verlag, 301–307.
Zapp A (1988). Automatische Straßenfahrzeugführung durch Rechnersehen, Dissertation, UniBwM / LRT
Dickmanns ED (2007). Dynamic Vision for Perception and Control of Motion. Springer-Verlag, London