Date of this Version
Most past and present work in machine perception has involved extensive static analysis of passively sampled data. However, it should be axiomatic that perception is not passive, but active. Furthermore, most past and current robotics research use rather rigid assumptions, models about the world, objects and their relationships. It is not so difficult to see that these assumptions, most of the time, in realistic situations do not hold, and hence, the robots do not perform to the designer's expectations.
Perceptual activity is exploratory, which implies probing and searching. We do not just see, we look. We do not only touch, we feel. And in the course, our pupils adjust to the level of illumination, our eyes bring the world into sharp focus, our eyes converge or diverge, we move our heads or change our position to get a better view of something, and sometimes we even put on spectacles.
Similarly, our hands adjust to the size of the object, to the surface coarseness and to the hardness or compliance of the material. This adaptiveness is crucial for survival in an uncertain, and generally, unfriendly world as millenia of experiments with different perceptual organizations have clearly demonstrated. Although no adequate account or theory of activity of perception has been presented by machine perception research, very recently, some researchers have recognized the value of actively probing the environment and emphasized the importance of data acquisition during the perception including head/eye movement.
Because of the realization of today's inadequacies of robotic performances, we in the GRASP laboratory at the University of Pennsylvania for the past five years have embarked on research in Active Perception and Exploratory Robotics. What follows is an expose of our theoretical foundation and some preliminary results. First, we shall describe what we mean by Active Perception, then we shall argue that Perception must also include manipulation, and finally, we will present Exploratory Robotics as a paradigm for extracting physical properties from an unknown environment.
Ruzena Bajcsy, "Active Perception and Exploratory Robotics", . August 1989.
Date Posted: 18 September 2007