The combination of computer vision for lateral guidance and a digital map for navigation has been previously explored by the authors on various private premises for the autonomous driving of road vehicles. The experiences gathered thereby has matured to the point when: experiments on public roads, in particular at highway exits and roundabouts, have become feasible and have actually been performed recently. The authors' experimental platform consists of an MB 609D van equipped with computer controlled steering, gas, brakes, and direction indicators as well as with a trinocular videocassette setup. Image sequences recorded by one computer selectable-camera from this trinocular setup are evaluated in parallel for lateral guidance as well as for the detection of intersections. Automatic longitudinal control was performed by evaluating the advance knowledge of road curvature from the digital road map in the navigation system Travelpilot of Robert Bosch GmbH, Stuttgart/Germany. The ability to select automatically that video camera which is expected to provide the optimum field of view for the actual phase of an automatic driving maneuvre turns out to provide significant advantages.
Multi-camera vision-based autonomous manoeuvring at road intersections
01.01.1994
648748 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Multi-Camera Vision-Based Autonomous Maneuvering at Road Intersections
British Library Conference Proceedings | 1994
|Autonomous vehicle parking and manoeuvring
Kraftfahrwesen | 1993
|Advances in Vision-Based UAV Manoeuvring Techniques
Springer Verlag | 2023
|