UAS planning and trajectory generation for safe and long-duration oceanic and coastal missions

Abstract

In this thesis a system that aims to extend the flight duration of small Unmanned Aerial Systems (UAS) is presented. The system was designed in the context of oceanic and coastal surveillance missions as part of the MarineUAS European project. Three main problems were identified: 1) the need to accurately estimate the wind field and the capability to identify features of interest, such as, wind shear, and gusts that may be suitable to allow energy extraction to improve flight duration. 2) the need to generate smooth trajectories that extract energy, considering the UAS platform dynamics and 3) the ability to follow such paths. For the first problem, the use of a direct computation method allows determining the wind field (wind velocity and wind rate of change) without the use of an optimal estimator. Nevertheless, different wind velocity estimation methods are compared, and the pros and cons of each are exposed; in addition, the identification of features is accomplished with a novel approach that performs a real-time statistical analysis of the distribution of the wind field estimates, allowing the characterization of the shear components and also any other potential features, like continuous and discrete gusts considering complex models that take into account not only the phenomena but the interactions with the ground and ocean through their respective boundary layers. For the second problem, a biomimetic approach is presented, replicating the trajectories of soaring birds by considering observations of these birds and the replication of their swooping maneuvers using smooth parametrized curves. This allows flexibility in the curve design and also the incorporation of dynamic constraints of the platform on it. The solution of the third problem takes into account the smooth curve that was generated and among it, a type 1 Bishop moving frame is generated. Then, a novel adaptive control method based on the vector-field theory approach is proposed to calculate the error equations and the respective control law, which permits the tracking of the designed trajectory for dynamic soaring. Furthermore, an additional step was added, in which the surveillance mission is re-configured on a waypoint-to-waypoint basis for a more efficient flight considering the identified wind field. The result was that the execution of soaring trajectories would not be executed during all the mission, but only in specific legs that fulfill specific characteristics.The primary goal was to design algorithms that implement these functions and to integrate these functionalities in a systems-engineering approach, in which the mission execution is the main priority. An extensive experimental campaign was performed at different levels, in which software-in-the-loop and hardwarein- the-loop tests, together with field tests, were executed to demonstrate the efficiency of the various functions separately and integrated. The field tests and the simulations consider different scenarios and UAS platforms, showing the performance of the system in different conditions. The results showed that the system could execute a more efficient mission, with savings of up to 20% in battery consumption, with the so-called of the Flight-Duration-Enhancement-System (FDES). Finally, the computational analysis showed that the system could be executed in real-time with minimum latency despite the use of sophisticated algorithms; this, together with the chosen software and hardware architectures allows portability to other hardware components and the possibility of incorporating additional functions.