Autonomy plays a key role for efficient operations in Urban Air Mobility (UAM) concepts, making precise localization of the aircraft crucial, especially during take-off and landing. Vertiports, designated landing sites for eVTOLs, can provide the means of ground-based precise positioning, such as augmented GNSS-based, vision-based or radar-based positioning. Among these, radar offers robustness against adverse weather and jamming. This paper proposes a loosely-coupled fusion architecture for both radar-based and GNSS-based 3D positioning with commonly used UAV sensors. The proposed architecture is based on Multiplicative Extended Kalman Filter (MEKF), which simplifies kinematic equations by incorporating a three-dimensional attitude error in the state vector while still maintaining a singularity- and gimbal-lock-free attitude representation with quaternions to describe the absolute attitude. Measurement covariances for radar and GNSS positions are calculated with the geometric dilution of precision (GDOP) values, which represent the geometric relationship between the error in distance measurements and the positioning error. Flight tests with an octocopter compare INS/GNSS, INS/Radar and INS/GNSS/Radar fusion schemes. Results demonstrate that incorporating radar alongside GNSS significantly improves positioning accuracy, while maintaining flexibility through a loosely-coupled fusion architecture.
Radar-Aided Navigation for Precise Approach at Vertiports
2025-06-18
753892 byte
Conference paper
Electronic Resource
English