Visual guidance of a small mobile robot using active, biologically-inspired, eye movements

Abstract

We introduce and implement a way of processing visual motion, which is directly inspired by the functional features of insects' visual systems. Rather than focusing on the computational performances of visual processing, alternative ways of detecting visual cues involving controlled eye movements are investigated. In this case, we show that the combination of retinal/eye movements significantly enhances the visual perception of motion for a sighted observer. The two kinds of eye movements introduced are: 1) pulsed retinal scanning, and 2) visually-guided eye rotations. Pulsed retinal scanning helps the eye to detect visual contrasts with enhanced reliability by increasing the S/N ratio of early visual signals. Rotations of the eye help a moving observer to maintain a continuous visual contact with obstacles while tracking their motion using pulses of retinal scanning. A direct demonstration of the performances has been implemented on a prototype eye, that successfully guides the motion of HECTOR, a small mobile platform, autonomously and in real-time.