Abstract
Localization is one of the most important tech-nologies needed to use Unmanned Aerial Vehicles (UAVs) inactual fields. Currently, most UAVs use GNSS to estimate theirposition. Recently, there have been attacks that target theweaknesses of UAVs that use GNSS, such as interrupting GNSSsignal to crash the UAVs or sending fake GNSS signals to hijackthe UAVs. To avoid this kind of situation, this paper proposes analgorithm that deals with the localization problem of the UAV inGNSS-denied environments. We propose a localization method,named as BRM (Building Ratio Map based) localization, for aUAV by matching an existing numerical map with UAV images.The building area is extracted from the UAV images. The ratioof buildings that occupy in the corresponding image frameis calculated and matched with the building information onthe numerical map. The position estimation is started in therange of severalkm2area, so that the position estimation canbe performed without knowing the exact initial coordinate.Only freely available maps are used for training data set andmatching the ground truth. Finally, we get real UAV images,IMU data, and GNSS data from UAV flight to show thatthe proposed method can achieve better performance than theconventional methods.
Abstract (translated)
URL
https://arxiv.org/abs/2008.01347