UAV Navigation in GNSS-Denied Environment Using Radio Beacons and Onboard Sensor Fusion
2024 (English)Independent thesis Advanced level (degree of Master (Two Years)), 28 HE credits
Student thesis
Abstract [en]
Positional knowledge is important for any aerial vehicle. The information enables both detailed surveillance and exploration of the surrounding environment. In recent decades, the usage and development of unmanned aerial vehicles (UAVs) has drastically increased in both civilian and military applications. Making them act autonomously is also of great interest, which is a considerable research field of its own. Precise positioning is especially important in this scenario. Currently, global navigation satellite systems (GNSS), e.g. the American global positioning system (GPS), are mostly used to acquire the positional information of UAVs in outdoor environments. However, these satellite-based systems are sensitive to occlusions and jamming. Also, the systems are mostly operated by nations which could be a potential vulnerability. Therefore, backup navigation systems are of great importance now and certainly in the future.
In this thesis, an outdoor UAV positioning system utilizing different sensors is proposed. More specifically, we employ radio beacons to measure angle of arrival (AOA) and received signal strength (RSS) from a transmitting UAV. Also, the onboard inertial measurement unit (IMU) and altimeter are used to provide further information of the dynamic behavior for robust estimation. First, a system design for tracking a UAV is proposed utilizing Kalman filters (KFs), where a decoupled tracker separating the orientation, using the IMU only, and position estimation is incorporated. Further, static positioning using nonlinear and linearized estimators are introduced for proper tracking filter initialization. Moreover, we develop a simultaneous localization and mapping (SLAM) framework to handle uncertainties in radio beacon position and orientation.
To evaluate the proposed positioning framework, extensive real-world experiments are conducted to analyze maximum measurement range and noise level of measurements of the radio beacons. Further, the system performance is verified through simulations with noise levels adopted from the hardware experiments. We compare the performance of different available measurements for position filter initialization and tracking of the UAV. Also, the performance of the proposed SLAM solution for the uncertain radio beacon states is investigated. The results show that RSS measurements can in most cases be omitted since the performance using AOA measurements is drastically better, enabling meter-like precision depending on the radio beacon separation. An onboard IMU and altimeter can both increase tracking performance and enable higher frequency information. The SLAM solution gives some promising results. However, it is not fully sufficient and more investigations are needed to fully solve the problem of uncertain radio beacon states.
Place, publisher, year, edition, pages
2024. , p. 88
Keywords [en]
sensor fusion, signal processing, uav, positioning, navigation, ekf, extended kalman filter, kf, kalman filter, imu, aoa, rss, received signal strength, slam, gnss-denied, gps-denied, ble, wsn, angle of arrival, aoa
National Category
Signal Processing
Identifiers
URN: urn:nbn:se:liu:diva-207051ISRN: LiTH-ISY-EX--24/5709--SEOAI: oai:DiVA.org:liu-207051DiVA, id: diva2:1893255
External cooperation
Swedish Defence Research Agency (FOI)
Subject / course
Electrical Engineering
Presentation
2024-08-27, Visionen Stora Konferensrummet, Linköping, 10:54 (English)
Supervisors
Examiners
Note
Opponent: William Nordström
2024-08-292024-08-292024-08-29Bibliographically approved