Banca de DEFESA: ÁLVARO PINTO FERNANDES DE NEGREIROS

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
STUDENT : ÁLVARO PINTO FERNANDES DE NEGREIROS
DATE: 04/08/2023
TIME: 09:00
LOCAL: Sala Virtual PPgEEC
TITLE:

High-level subsumption-based control architecture for sail-powered autonomous surface vehicles

KEY WORDS:

Sailboat, USV, Control Architecture, Subsumption Architecture, ROS, Gazebo, Reinforce Learning, PPO

PAGES: 150
BIG AREA: Ciências Exatas e da Terra
AREA: Ciência da Computação
SUBÁREA: Metodologia e Técnicas da Computação
SPECIALTY: Processamento Gráfico (Graphics)
SUMMARY:

We propose a high-level control architecture for an Autonomous Surface Vessel (ASV) designed to overcome the challenges associated with executing missions under varying weather conditions and with energy autonomy. The project is built on the understanding that while a number of low-level control techniques are readily available as open source, there remains a need for a high-level control architecture. This architecture would facilitate the creation of a resilient, versatile sailing vessel capable of handling any mission without requiring the user to master navigation specifics, naval procedures, or corner cases. The proposed solution involves developing a control architecture inspired by subsumption, and centered on hierarchical behaviors. This structure incorporates a variety of specialized behaviors adapted to different contexts, each of which is established using reinforcement learning techniques (PPO). A combination of Gazebo simulation environment with the ROS framework for training was used to validate the proposed architecture. This simulation enables the digital replication of the vessel's behaviors, which simplifies the implementation process and mitigates the challenges and costs tied to real-world sailing operations. The results of this study indicate that the high-level control architecture of the virtual sailing vessel was successful in passing both perimeter scanning and long-distance tests. This suggests that the ASV is equipped to navigate the several situations it might encounter in real-world missions.

COMMITTEE MEMBERS:
Presidente - 1345674 - LUIZ MARCOS GARCIA GONCALVES
Interno - 1242315 - PABLO JAVIER ALSINA
Externo à Instituição - EDUARDO CHARLES VASCONCELLOS - UFF
Externo à Instituição - DAVI HENRIQUE DOS SANTOS - UFRN
Externo à Instituição - ESTEBAN WALTER GONZALEZ CLUA - UFF
Externo à Instituição - JOAO MORENO VILAS BOAS DE SOUZA SILVA - IFRN