An adaptive localization based UAV framework for cooperative load transportation

<p dir="ltr"><b>Thesis Statement</b>: This thesis introduces a <b>novel push-based lift mechanism for autonomous cooperative UAV (uncrewed aerial vehicle) payload transport</b>, enhancing stability, efficiency, and control through adaptive localisation and aerodynamic optimisation. </p><p dir="ltr">Cooperative UAV systems rapidly evolve, enabling multiple drones to collaborate toward shared objectives. Multi-UAV platforms—particularly quadcopters with 3D vertical takeoff and landing (VTOL) capabilities—offer greater efficiency and robustness than single-agent systems. Equipped with onboard sensors, computation, and actuation, these platforms excel at tasks requiring precise positioning and coordinated load transport, with logistics, disaster response, and infrastructure inspection applications. However, realising safe and efficient cooperation remains challenging, particularly in synchronising flight dynamics and maintaining reliable, real-time communication in GPS-denied or degraded environments. </p><p dir="ltr">This thesis presents an adaptive framework for cooperative UAV-based load transportation, addressing critical localisation, coordination, and payload stabilisation challenges. It answers key research questions by developing and integrating subsystems spanning visual- inertial localisation, swarm coordination, modular actuation, and feedback- driven control. The framework is initially realised using commercial offthe- shelf (COTS) UAVs equipped with fiducial markers and onboard vision systems for reliable pose estimation. These insights inform the design of custom UAVs featuring onboard computing and advanced flight controllers, enabling real-time sensor fusion, trajectory planning, and closed-loop control in structured and GPS-denied environments. </p><p dir="ltr">The framework incorporates robust sensor fusion, combining visual odometry, SLAM, and inertial data to ensure stable navigation. A central contribution (Chapter 4) is developing a push-based cooperative lift mechanism. In contrast to conventional pullbased systems—often destabilised by aerodynamic disturbances—this design employs a Self-Balancing Tray (SBT) to minimise oscillations and enhance payload stability. This mechanical innovation combines synchronised actuation and control strategies (Chapter 5), enabling robust performance under dynamic conditions. </p><p dir="ltr">Chapter 6 introduces a decentralised communication and inter-UAV coordination architecture, enabling real-time trajectory sharing, adaptive role switching, and collision avoidance. This system ensures resilient multi-drone coordination even in GPScompromised settings. The complete framework is validated through extensive indoor and outdoor trials, demonstrating synchronised flight, stable load handling, and real-world applicability. Collectively, the framework advances multi-UAV cooperation by integrating localisation, mechanical design, control, and communication into a modular, resilient system for adaptive aerial transport.</p>