Air-Ground Cooperative Environment Modeling with Bounded Elevation Map

Abstract

UGVs' application are always restrained in complicated environment because of its limited perception ability. One of the solution is cooperation with UAV which has large perception scale. Thus, the fusion of the air-ground perceptive information is the key problem. Focused on this problem, the ESMF based 2.5D elevation modeling method is first proposed for its bound rather than probabilistic information is robust and reliable for path planning of UGVs. In the second, the fast fusion method between air and ground vehicles is designed, which achieve data fusion by updating the perception of UGV with UAV instead of intersecting the two observation sets. In the third the traversability is analyzed based on the elevation model described with bounded information, which shows the robust and reliable of the proposed method. In the last based on two typical limited perception ability: negative and ultrahigh obstacle an experiment is designed to verify the feasibility and validity of the propose method.