Integrating active perception with an autonomous robot architecture

Today's robotics applications require complex, real-time, high-bandwidth se nsor systems. Although many such systems have been developed, integrating t hem into an autonomous agent architecture remains an area of active researc h. We have integrated an active stereo vision system with an autonomous age nt architecture using a system of perceptual memory. Perceptual memory is a n important class of memory because it is designed for the 'behavior-based' portion of the agent's architecture, and not the deliberative portion. Thi s memory maintains current and recent task-dependent perceptual information , as well as expectations about the agent's immediate environment. Our syst em of perceptual memory is composed of visual primitives from our stereo sy stem, called proximity spaces. Each proximity space represents a virtual fo vea or locus of the agent's attention. As an application, we present a robo t that uses our system of perceptual memory and proximity spaces to 'attend to' multiple humans in a complex and unstructured indoor environment. (C) 1999 Elsevier Science All right reserved.