ace 
TOr 
the 
nes 
But 
ield 
de, 
ory 
otal 
our 
ing 
; in 
inal 
on- 
sual 
the 
was 
rror 
tion 
ork- 
intil 
ner. 
' ( 4 
ned 
urth 
turn 
ped. 
the 
com 
ond 
  
9. Conclusions 
The overall system architecture for flexible automation 
of vehicle guidance based on dynamic vision has been 
validated on two different testbeds. The basic sensory 
inputs to the system are video signals from CCD-TV 
cameras in combination with odometric and inertial sensor 
data from the vehicle body. The 4D-approach to dynamic 
vision yields a natural way to integrate different sources 
of sensor data. The visual navigation methods, path fol- 
lowing and landmark navigation, utilize the expectation- 
based approach perfectly by exploiting prediction error 
feedback with full spatio-temporal models for servo-main- 
taining an internal representation close to the real world 
objects. Both methods may be used in a complementary 
or an exclusive mode. In experimental results it has been 
shown that the combination of both yields very robust and 
precise navigation capabilities for an autonomous vehicle 
through structured environments. 
10. References 
[Bierman 75]: "Measurement Updating Using the U-D Factorization.’ 
Proc. IEEE Control and Decision Conf., Houston, TX, pp 
337-346, 1975. 
[Bierman 77]:Factorization Methods for Discrete Sequential 
Estimation, Academic Press, New York, 1977. 
[Brooks 87]: 'A Hardware Retargetable Distributed Layered 
Architecture for Mobile Robot Control." Proc. IEEE Robotics 
and Automation, Raleigh NC, April 1987, pp 106-110 
[Canny 86]: 'A computational Approach to edge detection.” IEEE 
Trans. Pattern Recog. Machine Intell., Vol. PAMI-8, No. 6, 
Nov.1986, pp 679-698. 
[Cox, Wilfong 90]: ’Autonomus robot vehicles’, Springer-Verlag, 
New York Berlin Heidelberg, 1990. 
[Dickmanns 88]: "Dynamic computer vision for mobile robot control.’ 
In: Jarvis, R.A.(ed.): Proceedings of the International Symosium 
and Exposition on Robots, Sydney, Nov. 1988, pp 314-327 
[Dickmanns, Christians 89]: 'Relative 3D-state estimation for 
autonomous visual guidance of road vehicles.” In: Kanada, T. 
e.a. (ed.): "Intelligent Autonomous Systems 2', Amsterdam, 
Dec. 1989, Vol.2 pp. 683-693; also appeared in: Robotics and 
Autonomous Systems 7 (1991), Elsevier Science Publ., pp 
113-123 
[Dickmanns, Graefe 88]:a) Dynamic Monocular Machine Vision, 
Machine Vision and Application 1, International Journal pp. 
223-240, Springer International, 1988. 
b) Applications of Dynamic Monocular Machine Vision, 
Machine Vision and Application 1, International Journal pp. 
241-261, Springer International, 1988. 
[Dickmanns, Mysliwetz 92]: ’Recursive 3D Road and Relative 
Ego-State Recognition.’ IEEE-Trans. PAMI, Special Issue on 
"Interpretation of 3D Scenes', 1992 . 
[Dickmanns, Zapp 86]: 'A Curvature- based Scheme for Improving 
Road Vehicle Guidance by Computer Vision.” In: "Mobile 
Robots’, SPIE Proc. Vol. 727, Cambridge, Mass., Oct. 1986, pp 
161-168 
[Dickmanns, Zapp 87]: Autonomous High Speed Road Vehicle 
Guidance by Computer Vision.” 10th IFAC Word] Congress, 
Munich, July 1987, Preprint Vol 4, pp 232-237 
[Graefe 90]: The BV V-Family of Robot Vision Systems, IEEE 
Workshop on Intelligent Motion Control, Istambul, Türkei, 
August 1990. 
[Hock 90a]:Dynamisches Modell für Dreiradfahrzeuge zur 
Landmarkennavigation, Universität der Bundeswehr München, 
UniBwM/LRT/WE13/IB90-4, März 1990 
[Hock 90b]:ATHENE, ein Projekt zur Landmarkennavigation, 6. 
Fachgesprách Autonome Mobile Systeme, Karlsruhe, November 
1990 
[Hock 91]:Landmark Navigation with ATHENE, 5. International 
Conference on Advanced Robotics, Pisa, Italy, June 1991 
[Kalman 60]:' A new Approach to Linear Filtering and Prediction 
Problems'. Trans ASME, Series D, Journal of Basic 
Engeneering, 1960, pp 35-40 
[Kuan, Phipps, Hsueh 86]: 'A real time road following vision system 
for autonomous vehicles'. Proc. SPIE Mobile Robots Conf., 
Vol. 727, Cambridge, MA, Oct. 1986, pp 152-160. 
[Kuhnert 85]: "Komponenten fuer die modellgestuetzte Interpretation 
dynamischer Szenen' (in German). Final Report of 
BMFT-Project O8IT 15113, Aerospace Dept., Universitaet der 
Bundeswehr Muenchen, 1985. 
[Kuhnert 90] : Dynamic Vision Guides the Autonomous Vehicle 
ATHENE, Japan-USA Symposium on Flexible Automation. 
Kyoto, Japan, July 1990. 
[Maybeck 79]:Stochastic Models, Estimation and Control, Vol. 1 
Academic Press, 1979 
[Müller 92]: 'Feedforward Control for Curve Steering for an 
Autonomous Road Vehicle.' IEEE Conf. on Robotics and 
Autmation, Nice, 1992. 
[Mysliwetz, Dickmanns 87]:'Distributed Scene Analysis for 
Autonomous Road Vehicle Guidance'. Proc. SPIE Conf. on 
Mobile Robots, Vol. 852, Cambridge, USA, 1987, pp 72-79 
[Mysliwetz 90]:"Parallelrechner-basierte Bildfolgeninterpretation zur 
autonomen Fahrzeugsteuerung'. Dissertation, Universitát der 
Bundeswehr Muenchen, LRT, 1990. 
[Turk, Morgenthaler, Gremban, Marra 87]:' Video road-following for 
the autonomous land vehicle'. Proc. IEEE Int. Conf. Robotics 
and Automation, Raleigh, NC, Apr. 1987, pp 273-280. 
Wallace et al. 86]:'Progress in Robot Road-Following'. Proc. Int. 
Conf. Robotics and Automation, San Francisco, CA, Apr. 1986, 
pp 1615-1621. 
[Wuensche 86]: "Detection and Control of Mobile Robot Motion by 
Real-Time Computer Vision'. In Marquino N. (ed), Advances in 
Intelligent Robotics Systems. Proc. of the SPIE, Vol. 727, 
Cambridge. Mass., 1986, pp 100-109. 
[Wünsche 88]:Bewegungssteuerung durch Rechnersehen. 
Fachberichte Messen, Steuern Regeln Bd. 20, Springer-Verlag, 
Berlin, 1988 
[Zapp 88]: Automatische StraBenfahrzeugführung durch 
Rechnersehen, Dissertation an der Fakultát für Luft- und 
Raumfahrttechnik der Universität der Bundeswehr München, 
1988. 
[Zimdahl, Rackow, Wilm 86]: 'OPTOPILOT - ein Forschungansatz 
zur Spurerkennung und Spurfithrung bei StraBenfahrzeugen. 
VDI Berichte Nr.162, 1986, pp 49-60