Geometric Control and Estimation for Autonomous UAVs in Ocean Environments
Open AccessThis dissertation presents the development of a multi-rotor UAV platform capable of operating in ocean environments where the base of operation is a surface vessel. The rolling and pitching motion of the ship, coupled with the turbulent wind conditions in ocean, makes autonomous flight of UAV in such environments challenging. To address that, a novel estimator and a geometric controller for UAV are developed, and they are implemented on a custom-made UAV hardware and software platform. Further, a visual-inertial odometry approach is integrated into the developed system to operate UAV in GPS-denied environments.First, a state estimator for UAV is developed to address time delay in sensor measurements. Autonomous UAV are equipped with various sensors, where a state estimator integrates the sensor measurements with the dynamics to estimate its state. The onboard sensors have distinct characteristics. For example, GPS provides position measurements at a low frequency with inherent processing delay, while a gyro measures the angular velocity frequently without any delay, but with a time-varying bias. To address this, an extended Kalman filter is designed to handle measurement delays where only a subset of measurements is delayed by a fixed duration. This is critical for autonomous UAV in ocean environments where a low-cost realtime kinematics GPS is utilized for high-precision position measurements. Second, a novel adaptive geometric controller is constructed with decoupled-yaw controls. Multirotor UAVs are often underactuated as it can generate thrust along a particular direction fixed to its body. As such, the UAV has to rotate its body to change the direction of the thrust. However, any rotation of the body about the axis of thrust, referred to as yawing, does not have any effect on the resultant thrust, and therefore, it is decoupled from the translational dynamics. It is shown that by splitting the attitude control of the multirotor UAV into the yawing dynamics and the remaining rolling and pitching dynamics, the position tracking performance can be improved. Further, the proposed controller is augmented with adaptive control terms to handle unmodeled dynamics and external disturbances such as wind. Third, a custom UAV platform is developed to verify the above estimator and controller in various indoor and outdoor environments. Despite the popularity, it is cumbersome to implement advanced, exploratory estimation and control schemes to open-source flight control systems, such as Pixhawk. Here, both a hardware platform and a flight software are developed such that any innovative estimation and control approaches can be tested easily. The hardware system is based on a general-purpose computing module powered by GPU, and it further includes sensor integration and actuation using custom circuit boards, water-tight electronics, and an active heat dissipation system. The flight software is built with multi-threads programming such that various tasks of control, estimation, and communication are executed simultaneously. This is tested with real-time flight experiments in indoor, outdoor, and ocean environments.Finally, an open-source visual-inertial odometry (VIO) solution is integrated with the above UAV system for localization in GPS-denied environments. VIO is executed onboard along with the proposed estimator and the controller, while accounting the time-delay caused by image processing. This is verified with autonomous landing of a multirotor UAV on the flight deck of a Navy research vessel in Chesapeake Bay.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.
| Thumbnail | Title | Date Uploaded | Visibility | Actions |
|---|---|---|---|---|
|
|
Gamagedara_gwu_0075A_16055.pdf | 2022-10-04 | Open Access |
|