A cooperative positioning approach of Unmanned Aerial Vehicles with Accuracy and Robustness

Efficient localization is crucial for the autonomous navigation of Unmanned Aerial Vehicles (UAVs) in GPS-denied environments. Current onboard visual positioning approach may easily deteriorate with varying illumination and unpredicted obstacles. To tackle this issue, a novel cooperative positioning approach is proposed in this paper. This approach is comprised of a marker-based position estimation and an end-to-end UAV detection. For the marker-based method, a prior known pattern is attached on UAV and the Perspective-n-point (PnP)algorithm is implemented to provide an accurate estimation of the UAV's position. Meanwhile, to enhance the localization robustness, an end-to-end detection method is designed based on FCOS neural network and ultra-wideband (UWB) range measurement. Then, with an additional data fusion, the aforementioned two methods are integrated eventually. In this way, the accuracy and robustness of the localization can be enhanced accordingly. Extensive experiments under different environments are carried out to verify the effectiveness of the proposed method. Results demonstrate that the proposed method can estimate the flight trajectory of UAV with high accuracy and robustness in cluttered environments with varying illumination.