An integrated computational model of an autonomous, visually guided undersea vehicle

This paper reports progress made on the development of a computational model of a visually guided autonomous underwater vehicle. The model is based on knowledge obtained from laboratory studies in psychophysics and neurophysiology as well as on new developments in computer science. It is embodied in a computer simulation of a predatory SWIMMER-an underwater creature capable of finding food in its simulated undersea world on the basis of an excellent understanding of its visual and spatial environment. Brief discussions of the visual image processing, decision making, and navigational aspects of the simulation are presented. The work can be considered to be either a psychological theory of visual perception or an initial step leading to functional programming for autonomous visual control of an underwater vehicle