Bio-Inspired Stereoscopic Ranging Imager for Robot Obstacle Avoidance

This thesis presents and evaluates a bio-inspired vision system design to increase the depth field of a stereoscopic ranging imager. Two key attributes of the human vision system are leveraged in this design. The first attribute is image stabilization similar to the inner ear semi-circular canals and neck muscles. To accomplish this, an accelerometer and servos were used to stabilize the imager platform. The second human vision attribute used by the design is the ability to change the focal vector. This is accomplished by a servo that tilts the imagers in unison and separate servos that enable each imager to pan independently. The performance metrics of depth field size and resolution of this method are compared to a system that has statically-mounted imagers and a fixed video platform.