Drift-Free 4DoF Pose Estimation for UAVs: Minimizing UWB Devices With IMU Integration

Accurate and robust pose estimation is paramount for uncrewed aerial vehicles (UAVs) across a spectrum of applications, including indoor navigation and autonomous landing. Traditional methods predominantly rely on ultrawideband (UWB) technology for position estimation and inertial measurement units (IMUs) for orientation estimation. However, IMU-based heading estimates are susceptible to drift and magnetic interference, while existing orientation-coupled UWB systems require extensive deployments of tags and anchors, leading to high costs and limited scalability. This study addresses the challenge of achieving a drift-free six-degree-of-freedom (6-DOF) pose estimation with minimum UWB devices by exploiting the drift-free roll and pitch angles provided by the IMU to constrain the problem to four degrees of freedom (4-DOF), corresponding to 3DoF position and 1DoF heading angle (yaw). A key contribution is the rigorous derivation of the minimal sensor configuration, demonstrating that only two UWB tags and two anchors are sufficient for simultaneous position and heading estimation, significantly reducing hardware requirements. Additionally, we introduce two novel metrics based on the Cramér-Rao lower bound (CRLB): position dilution of precision (P-DOP) and heading dilution of precision (Heading-DOP). These metrics quantify the impact of distance measurement errors on localization estimation errors for 4-DOF pose estimation based solely on range measurements. We analytically derive the theoretical lower bounds for these metrics in closed form and explicitly define the anchor placement conditions required to achieve optimal performance. Building on these insights, we propose a simple yet effective multiobjective optimization strategy to design anchor configurations that minimize both position and heading estimation errors, balancing real-world tradeoffs. Extensive simulations and real-world experiments validate the effectiveness and practicality of the proposed approach. Experimental results demonstrate that the optimal anchor configuration achieves a position RMSE of 0.083 m and a yaw RMSE of 2.40°, both surpassing the other methods compared. Our codes are open-sourced at https://github.com/LvJohny/dual-uwb-4dof-pose.git