Abstract:
A model of event driven anticipatory learning is described and applied to a number of attention situations where one or several visual targets need to be tracked while being intermittently occluded. The model combines covert tracking of multiple targets with overt control of a single attention focus. The implemented system has been applied to both a simple scenario with a car that is occluded in a tunnel and a complex situation with six simulated robots that need to anticipate the movements of each other. The system is shown to learn very quickly to anticipate target movements. The performance is further increased when the simulated robots are allowed to cooperate in the tracking task.
