A COLREGS-compliant collision avoidance system for unmanned surface vehicles

• Ano de defesa: 2020
• Autor(a) principal: Jurak, Darlan Alves
• Orientador(a): Amory, Alexandre de Morais
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Pontifícia Universidade Católica do Rio Grande do Sul
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação
• Departamento: Escola Politécnica
• País: Brasil
• Palavras-chave em Português: Robotic Operating System (ROS); Normas da Autoridade Marítima (NORMAM); Evasão de colisão
• Palavras-chave em Inglês: Unmanned Surface Vehicle (USV); Convention on the International Regulations for Preventing Collisions at Sea; Collision Avoidance (COLREGS); Artificial Terrain Cost A*
• Área do conhecimento CNPq: CIENCIA DA COMPUTACAO::TEORIA DA COMPUTACAO
• Link de acesso: http://tede2.pucrs.br/tede2/handle/tede/9471

Unmanned surface vehicles (USVs) are a category of aquatic robots that act unmanned on the water surface, performing autonomous behavior or being remotely controlled. In recent decades, several studies have been carried out to make USVs autonomous. The main challenges are related to collision avoidance, accurate navigation on the high seas, and compliance with international maritime rules, such as the Convention on the International Regulations for Preventing Collisions at Sea (COLREGS). Current applications of USVs include monitoring the environment, exploiting ocean resources such as oil and gas, port and coastal surveillance for military purposes, transport, and scientific research. COLREGS determines rules that must be followed by sailors to avoid collisions in possible collision scenarios, such as crossing, head-on, and overtaking. Currently, direct collisions between vessels represent 60% of accidents at sea, and the violation of COLREGS causes 56% of collisions. Therefore, USVs must comply with COLREGS. In this master’s thesis, we present the system we developed to guide USVs on autonomous missions while following the COLREGS. The main guidance module of our system is a path planner that follows the COLREGS. We adapted the Artificial Terrain Cost method, presented by Agrawal [3] et al., to our solution, blocking locations in the search space that are not compliant with COLREGS, through the creation of virtual obstacles. We integrated our system to a differential boat available in the USV_sim simulator, a simulator for USVs, capable of generating realistic disturbances, such as wind and water current. To evaluate the proposed system, we measure the computational time, the maximum sustained wind, and the minimum distance maintained when encountering another vessel in the scenarios of head-on, crossing from the left, crossing from the right, and overtaking encounters.