New Stereovision for Human-Robot Communications

Better communications between robots and humans requires an understanding of human senses, and at the same time, the ability to express these senses. Robot functions must be able to approximate as much as possible the five human senses of sight, hearing, smell, taste and touch. Further, the system in which these five senses are integrated must be understood. The expression of such senses by robots should not be confined to merely simulating human behaviors but needs to address unknown and difficult problems related to the manifestation of the inner aspects of human mentality. In this paper, we report on a new artificial vision system that is similar to the human sense of sight, with the target of realizing functions for understanding the processes of human sight. First, we introduce the new artificial visual sensing system and explain problems of virtual images and occlusion, blind spots in conventional binocular stereovision systems. We then discuss a technique for solving these problems that utilizes the new functions of the artificial vision system we developed and introduce a prototype of the new system and its configuration.