One of the goals of rescue robotics has been to allow machines to sense and act where people cannot go. Deployed with various sensors and under the guidance of trained operators these systems have shown great potential for finding victims in rubble. However, the current state of mobility of all deployed rescue robots is such that they can only be employed in very constrained circumstances where surfaces are mostly regular and acceptable for various modes of tracked and wheeled mobility. We have been exploring an alternative employing the existing abilities of trained US.4R canines to search nibble both quickly and effectively. This paper discusses one of the potential areas where canines might be augmented to deliver the desired sensing and remote control of a robot.