The Robotics Institute at CMU has released a new technical report, titled Terrain Classification from Aerial Data to Support Ground Vehicle Navigation (PDF format), detailing a method of improving navigation for high-speed off-road ground robots by providing additional sensor data from an unmanned flying robot. CMU has actually developed a working system, called the NREC PerceptOR, comprised of autonomous UGV (the PerceptOR) and an autonomous helicopter (the flying eye). By working together as a team, the robots are able to find the optimal path for the ground vehicle and avoid dangerous terrain such as cliffs that would not normally be visible to the UGV. Not suprisingly, the research is being funded by DARPA.