The paper describes an approach to an optimal gaze control system for autonomous vehicles. This gaze control system should not only determine the viewing direction ad hoc for the present moment, but also plan and optimize the viewing behavior in advance for a certain period of time. For planning the viewing behavior the situation must be predicted. The expression 'situation' includes not only the physical situation, but also the so-called perceptive situation and subjective situation. The perceptive situation takes the present perception status, the present knowledge and the perception capabilities of the autonomous vehicle into consideration. The subjective situation contains locomotion tasks, status of vehicle sensors and actuators and subjective meanings of external objects resulting from the own point of view. For simulating the situation, the states and variances of all objects must be predicted. In the system for EMS-Vision, parts of the gaze control unit are already implemented and tested in both UBM test vehicles VaMoRs and VaMP. The embedding of the gaze control unit in the EMS-Vision architecture is discussed.
[1]
E. D. Dickmanns,et al.
EMS-Vision: enhanced abilities for locomotion
,
2000,
Proceedings of the IEEE Intelligent Vehicles Symposium 2000 (Cat. No.00TH8511).
[2]
E.D. Dickmanns,et al.
EMS-vision: recognition of intersections on unmarked road networks
,
2000,
Proceedings of the IEEE Intelligent Vehicles Symposium 2000 (Cat. No.00TH8511).
[3]
Dana H. Ballard,et al.
Principles of animate vision
,
1992,
CVGIP Image Underst..
[4]
Ernst D. Dickmanns,et al.
An Expectation-based, Multi-focal, Saccadic (EMS) Vision System for Vehicle Guidance
,
2000
.
[5]
M. Pellkofer,et al.
EMS-Vision: a perceptual system for autonomous vehicles
,
2000,
Proceedings of the IEEE Intelligent Vehicles Symposium 2000 (Cat. No.00TH8511).
[6]
E. D. Dickmanns,et al.
EMS-Vision: mission performance on road networks
,
2000,
Proceedings of the IEEE Intelligent Vehicles Symposium 2000 (Cat. No.00TH8511).