Implementation and validation of obstacle-avoidance and navigation control strategies for fully-actuated multirotors

Fully actuated multirotors offer independent control of translational and rotational motion, enabling stable maneuvers and advanced autonomous navigation capabilities. Among these platforms, the tilted hexarotor provides full six degrees of freedom (6-DoF) actuation, allowing complex trajectory execution without the coupling limitations of conventional multirotors. However, exploiting these capabilities requires advanced control strategies capable of maintaining accurate trajectory tracking while handling nonlinear Unmanned Aerial Vehicles (UAV) dynamics and obstacle avoidance requirements. This thesis presents the design, implementation, and comparative evaluation of three control architectures for a fully actuated tilted hexarotor: a Standard PID Controller (SC), a Geometric Controller (GC), and a Flatness-Based Controller (FC). The controllers are implemented and evaluated in MATLAB®/Simulink® through trajectory-tracking tests using both quantitative performance metrics and qualitative analysis. Based on the obtained results, the FC is selected as the most suitable architecture due to its high tracking accuracy, lower control effort, and compatibility with reference-based obstacle avoidance integration. Two obstacle detection and avoidance (ODA) strategies are subsequently developed and integrated with the selected controller: a virtual proximity-sensor-based arc avoidance method and an Artificial Potential Field (APF) approach. Both schemes are validated under different obstacle configurations. The virtual-arc strategy provides predictable geometric clearance and smooth avoidance trajectories but introduces noticeable transients in the velocity and acceleration responses at arc transitions. In contrast, the APF-based strategy generates smoother avoidance maneuvers while maintaining accurate tracking performance. Consequently, the APF-based ODA is selected as the preferred solution. Finally, the FC + APF-based ODA is validated within MATLAB®/Simulink® environment using Simscape™ Multibody™ and the RotorSuite toolbox. Comparison between the analytical and physics-based models shows only minor discrepancies, confirming the validity of the proposed navigation framework and demonstrating its feasibility for autonomous operation in cluttered environments.