A.3.7 Turning off onto an unknown crossroads

A.3.7 a) Turn right on two-lane roads with lane markings

From GPS- and map information and from the sequence of mission elements to be performed, attention is controlled to detecting a cross-road on the proper side of the road driven; left or right turns require different procedures depending also on the width of the road driven.

The following generic capabilities have been developed in Hardware-In-the-Loop simulation

H.4.1 HIL-Sim RoadVehGuidance

and later on proven by real-world experiments with VaMoRs on our test track on the former airport Neubiberg as well as on the test site Pfullendorf:

  • Detection of a crossroad within a certain range of intersection angles at sufficiently large ranges.

  • Determine range to the intersection point, width of the crossroad and the intersection angle by active vision while approaching.

  • Selection of the proper feed-forward turn-off steering program over time and of the optimal trigger point for initiation.

  • Superposition of lateral feedback control derived from crossroad tracking in the final part of the stereotypical maneuver to counteract perturbations and errors accumulated.

Video 23 TurnoffNbbTaxiway 1996

A.3.7 b) Turn left on a dirt road without lane markings exploiting trinocular

H.1.3 EMS vision system

Top row: Saccadic perception of cross road and intersection at large range with the tele-camera; each color image with a small field of view is taken at a different time representing the extreme directions (left and right) during the saccade. The center image taken during saccadic motion is not evaluated; in the other two the search paths for edge extraction are marked as red lines.

Center rows a) to c): Left and right wide-angle images of MarVEye while performing the turn at the intersection; in a) the vehicle is still driving ~ in the direction of the old road while looking ‘over the shoulder’ to the left with the far part of the crossroad appearing in the lateral small window. The vehicle starts turning and centers gaze onto the crossroad [b) and c)].

Row d): Again, one tele image is shown (left) while in the lower right the red arrow on the blue track shows the path driven.

Video 30 TurnLeftSacc DirtRoad 2001


Baten S, Müller N. (1995). Navigation with an Autonomous Vehicle on Arbitrary Roads. In Linkwitz et al. (eds): High Precision Navigation. Dümmler Verlag, Bonn, pp. 415-42

Dickmanns ED, Müller N (1995). Scene Recognition and Navigation Capabilities for Lane Changes and Turns in Vision-Based Vehicle Guidance. Control Engineering Practice, 2nd IFAC Conf. on Intelligent Autonomous Vehicles-95, Helsinki

Müller N, Baten S (1995). Image Processing based Navigation with an Autonomous Car. Int. Conf. on Intelligent Autonomous Systems, IAS-4, Karlsruhe, pp 591-598

Müller N (1996). Autonomes Manövrieren und Navigieren mit einem sehenden Straßenfahrzeug, Diss., UniBwM / LRT

Lützeler M, Dickmanns ED (2000). EMS-Vision: Recognition of Intersections on Unmarked Road Networks. Proc. Int. Symp. on Intelligent Vehicles (IV’2000), Dearborn, (MI)

Gregor R, Lützeler M, Pellkofer M, Siedersberger KH, Dickmanns ED (2001). A Vision System for Autonomous Ground Vehicles with a Wide Range of Maneuvering Capabilities. Proc. ICVS, Vancouver

Gregor R (2002). Fähigkeiten zur Missionsdurchführung und Landmarkennavigation. Dissertation, UniBwM / LRT

Pellkofer M, Hofmann U, Dickmanns ED (2003). Autonomous cross-country driving using active vision. SPIE Conf. 5267, Intelligent Robots and Computer Vision XXI: Algorithms, Techniques, and Active Vision. Photonics East, Providence, Rhode Island

Pellkofer M (2003).  Verhaltensentscheidung für autonome Fahrzeuge mit Blickrichtungssteuerung. Diss., UniBwM, LRT

Siedersberger K-H (2004).  Komponenten zur automatischen Fahrzeugführung in sehenden (semi-) autonomen Fahrzeugen. Dissertation, UniBwM, LRT, 2004

Dickmanns ED (2007). Dynamic Vision for Perception and Control of Motion. Springer-Verlag, (Chapter 10)