Study on direct perception of collision avoidance mediated by brightness differences

Generally, ecological approaches toward obstacle avoidance employ "optic flow" as visual information. However, it is difficult to apply those approaches to navigations of a robot in the "real world", since optic flow is obtained through a complex image processing. In this paper, we proposed a method based on brightness differentials which are obtained by a simple process. Using brightness differentials as an intrinsic metrics, robots perceive "affordance" from obstacles and can avoid them with the least effort.