Investigation of different Kalman filtering architectures for navigation in UAV swarms

Date

2025-8

Author

Turan, Sena

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

516
views

685
downloads

Cite This

This thesis investigates navigation in unmanned aerial vehicle (UAV) swarms operating in GNSS-denied environments, with a focus on how flight paths, swarm size, sensor quality, and filter architecture affect estimation accuracy. Both local and master Kalman Filter (KF) configurations, as well as a decentralized local KF approach, are analyzed under different scenarios, including nominal conditions and communication disruptions. The swarm architecture consists of a parent UAV equipped with a high-quality inertial measurement unit (IMU) and optionally a CCD camera, along with multiple child UAVs using lower-cost tactical-grade IMUs. Communication is maintained through radio frequency (RF) datalinks, which enable the exchange of navigation data. Results show that when only range measurements are available, the master KF significantly improves position accuracy over local KFs, although performance remains limited without attitude corrections. With additional position and attitude measurements, the master KF offers only marginal improvements, while reliable reference data from the parent UAV’s high-quality IMU proves essential for stable estimation. In the decentralized configuration, child UAVs run their own local filters using position and attitude measurements sent by the parent UAV. This structure performs robustly under nominal conditions, and when RF datalink outages occur, the integration of a chi-square test allows faulty measurements to be rejected, maintaining errors close to nominal levels. Overall, the findings emphasize the importance of reliable reference data, the value of attitude measurements for improving accuracy, and the advantages of decentralized filtering in handling communication faults. These insights contribute to developing scalable and fault-tolerant swarm navigation strategies for surveillance, search and rescue, and cooperative transport.

Subject Keywords

Swarm Navigation, Unmanned Aerial Vehicle (UAV), Kalman Filter (KF), Sensor Fusion, Position Estimation

URI

https://hdl.handle.net/11511/115686

Collections

Graduate School of Natural and Applied Sciences, Thesis