Learning-based Control of Perception for Mobility

Machine perception plays an important role in any intelligent system, and in particular, guiding an autonomous mobile agent. Machine perception techniques have progressed significantly in recent years, however perception systems are still plagued by a lack of flexibility and an inadequacy in performance speed for use in real-time tasks. To overcome these problems, we have applied integrated karning lechniques to a perception system that is based on a selective sens ingpad ip . The incorporation of multiple learning algorithms at different levels in our perception system provides a great deal of flexibility and robustness when performing different perceptual tasks. Making use of a selective sensing paradigm allows the system to eliminate a large amount of non-pertinent sensory data so that processing speed is greatly increased. We are implementing such a perception system to be used on an autonomous mobile agent. In this paper, we describe our methodology and give a preliminary example of learning within our perception system.