Disturbance-Aware Motion Planning and Control of Unmanned Aerial and Surface Vehicles

QC 20240826

Abstract

This thesis concerns motion planning and control of underactuated unmanned aerial and surface vehicles with special attention to disturbances.  In the first part of the thesis, we examine trajectory tracking using Prescribed Performance Control (PPC) for the classes of underactuated aerial and surface vehicles, assuming that the model parameters are unknown. Due to the underactuation, the original PPC methodology is redesigned to accommodate the specifics of the considered underactuated dynamical systems. We prove the stability of the proposed control schemes and support them with numerical simulations on the quadrotor and boat models. Furthermore, we propose enhancements to the Kinodynamic Motion Planning via Funnel Control (KDF) framework. The kinodynamic motion planning is based on the Rapidly-exploring Random Trees (RRT) algorithm, and our improvements are in the optimization-based generation of smooth, collision-free trajectories using B-splines.Real-world experiments were conducted for the surface vehicles and tested the advantages of the proposed enhancements to KDF. The second part of the thesis is devoted to the rendezvous problem of the autonomous landing of a quadrotor on a boat based on distributed Model Predictive Control (MPC) algorithms. We propose an algorithm that assumes a minimal exchange of information between the agents, which is the rendezvous location, and an update rule to maintain the recursive feasibility of the landing. Moreover, a convergence proof is presented without enforcing the terminal set constraints.  Finally, we investigate a leader-follower framework and present an algorithm for multiple follower agents to land autonomously on the landing platform attached to the leader. An agent is equipped with a trajectory predictor to handle the cases of communication loss and avoid inter-agent collisions. The algorithm is tested in a simulation scenario with the simultaneous landing of multiple agents. %and a real-world scenario of the detect-and-avoid problem with two UAVs.

In the third part of the thesis, we examine the usage of the disturbance models and methods to refine them using available data and sensory measurements iteratively.Contraction-based control methods enable safety guarantees for unmanned aerial vehicles with respect to the disturbance-aware plans generated by MPC augmented with disturbance models. Disturbance models are inferred by data-driven identification and learning and further refined using adaptive control methods.The exploration-exploitation algorithm is presented for learning previously unseen disturbances.Finally, the framework is tested in a simulation scenario of the autonomous landing of a UAV on a surface vehicle.

urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-352181