UWB INTEGRATION WITH IMU SENSORS: ADVANCING PRECISION POSITIONING

Accurate and robust indoor localization is a critical enabler for Internet of Things (IoT) applications, particularly for the autonomous navigation of Unmanned Aerial Vehicles (UAVs). This thesis addresses the distinct limitations of two primary positioning technologies: Ultra-Wideband (UWB), which provides high-precision absolute positioning but is susceptible to significant high-frequency noise , and Inertial Measurement Units (IMUs), which offer smooth, high-rate relative motion data but suffer from cumulative drift over time. This work develops and evaluates a sensor fusion framework that integrates UWB and IMU data to produce a single, reliable position estimate. A Kalman Filter, built on a 6-dimensional state vector (3D position and 3D velocity), is implemented. The filter's prediction step utilizes the drone's onboard IMU acceleration data to project a physically smooth trajectory , while the update step uses absolute position measurements from the UWB system to correct for drift. The framework's performance was experimentally validated using DJI Tello drones , with a Vicon motion capture system serving as the high-fidelity ground truth. The study first demonstrates the critical impact of anchor geometry on raw UWB performance, establishing that non-coplanar, 3D anchor configurations are essential to minimize the Geometric Dilution of Precision (GDOP) and ensure reliable altitude (Z-axis) estimation. The results show that the implemented Kalman Filter successfully fuses the complementary strengths of both sensors. It effectively rejects UWB noise and eliminates IMU drift , producing a fused trajectory that is dramatically smoother than the raw UWB data and accurately tracks the Vicon ground truth. This hybrid system provides a robust, drift-free, and precise positioning solution suitable for closed-loop flight control in dynamic indoor environments. This work develops and evaluates a sensor fusion framework that integrates UWB and IMU data to produce a single, reliable position estimate. A Kalman Filter, built on a 6-dimensional state vector (3D position and 3D velocity), is implemented. The filter's prediction step utilizes the drone's onboard IMU acceleration data to project a physically smooth trajectory , while the update step uses absolute position measurements from the UWB system to correct for drift. The framework's performance was experimentally validated using DJI Tello drones , with a Vicon motion capture system serving as the high-fidelity ground truth. The study first demonstrates the critical impact of anchor geometry on raw UWB performance, establishing that non-coplanar, 3D anchor configurations are essential to minimize the Geometric Dilution of Precision (GDOP) and ensure reliable altitude (Z-axis) estimation. The results show that the implemented Kalman Filter successfully fuses the complementary strengths of both sensors. It effectively rejects UWB noise and eliminates IMU drift , producing a fused trajectory that is dramatically smoother than the raw UWB data and accurately tracks the Vicon ground truth. This hybrid system provides a robust, drift-free, and precise positioning solution suitable for closed-loop flight control in dynamic indoor environments.