A Modular Sensor Fusion and Quality-Driven Sensor Selection Algorithm for UAV Navigation

Unmanned Aerial Vehicles (UAVs) are increasingly integrated in various applications. The urgent need for accurate navigation algorithms is emphasized in various applications. In this paper, we propose an integrated framework focusing on robust attitude and position/velocity estimation for UAV navigation using Kalman filtering employing Kalman filtering (KF) technique and chi-square test based sensor fusion quality assessment. Our approach integrates sensor data through a filter-based approach, and Kalman filter handles attitude estimation and position/velocity estimation. The modal architecture enhances robustness and facilitates seamless integration of multiple sensors while prioritizing reduced attitude estimation and position/velocity estimation. Accurate attitude estimation is also prioritized to reduce the risk in case of deviation. Accurate attitude estimation is also prioritized to reduce the risk of deviation in case of higher order state estimators. Experimental validation on the “SHENG PROUAV” model proves the effectiveness of our framework and demonstrates accurate attitude estimation results.