Closing the gap between simulation and reality in the sensor and motion models of an autonomous AR.Drone

This article describes a method to develop an advanced navigation capability for the standard platform of the IMAV indoor competition: the Parrot AR.Drone. Our development is partlybased on simulation, which requires both a re-alistic sensor and motion model. This article de-scribes how a visual map of the indoor environ-ment can be made, including the effect of sensor noise. In addition, validation results for the mo-tion model are presented. On this basis, it shouldbe possible to learn elevation maps, optimal paths on this visual map and to autonomously avoid obstacles based on optical flow.