Bounded Active Perception
Résumé
There are two kinds of perception : active and passive. This paper is an attempt to take advantage of active perception to improve the agent's perception of relevant information. Through the data filtering capacity, active perception is a useful tool for modeling human-like bounded perception. Using such filters, either the agent or the environment take an active role. We determine several unsolved issues in active perception and do several proposals to implement our concept on the active perception framework. To this end, we benefit from the cognitive science studies. Finally, our proposals are applied to the case of multiagent traffic simulation.