Ultra-Wideband Based Pose Estimation for Small Unmanned Aerial Vehicles

This paper proposes a 3-D local pose estimation system for a small Unmanned Aerial Vehicle (UAV) with a weight limit of 200 g and a very small footprint of 10 cm<inline-formula> <tex-math notation="LaTeX">$\times$ </tex-math></inline-formula> 10 cm. The system is realized by fusing 3-D position estimations from an Ultra-Wide Band (UWB) transceiver network with Inertial Measurement Unit (IMU) sensor data and data from a barometric pressure sensor. The 3-D position from the UWB network is estimated using Multi-Dimensional Scaling (MDS) and range measurements between the transceivers. The range measurements are obtained using Double-Sided Two-Way Ranging (DS-TWR), thus eliminating the need for an additional clock synchronization mechanism. The sensor fusion is accomplished using a loosely coupled Extended Kalman Filter (EKF) architecture. Extensive evaluation of the proposed system shows that a position accuracy with a Root-Mean-Square Error (RMSE) of <inline-formula> <tex-math notation="LaTeX">$0.20\,cm$ </tex-math></inline-formula> can be obtained. The orientation angle can be estimated with an RMSE of 1.93°.