Traversability: A Case Study for Learning and Perceiving Affordances in Robots

The concept of affordances, introduced in psychology by J. J. Gibson, has recently attracted interest in the development of cognitive systems in autonomous robotics. In earlier work (Sahin, Cakmak, Dogar, Ugur, & Ucoluk), we reviewed the uses of this concept in different fields and proposed a formalism to use affordances at different levels of robot control. In this article, we first review studies in ecological psychology on the learning and perception of traversability in organisms and describe how the existence of traversability was judged to exist. We then describe the implementation of one part of the affordance formalism for the learning and perception of traversability affordances on a mobile robot equipped with range sensing ability. Through experiments inspired by ecological psychology, we show that the robot, by interacting with its environment, can learn to perceive the traversability affordances. Moreover, we claim that three of the main attributes that are commonly associated with affordances, that is, affordances being relative to the environment, providing perceptual economy, and providing general information, are simply consequences of learning from the interactions of the robot with the environment.