Self-organized flocking control for micro aerial vehicles swarm in global navigation satellite system-denied environments

• Ano de defesa: 2021
• Autor(a) principal: Amorim, Thulio Guilherme Silva de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Informática; Programa de Pós-Graduação em Informática; UFPB
• Palavras-chave em Português: Robótica de enxame; Movimento coordenado; Micro veículos aéreos; Veículos aéreos não-tripulados; Auto-organização; Swarm robotics; Flocking; Micro aerial vehicles; Unmanned aerial vehicles; Self-organization; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/23125

Flocking, also known as coordinated motion, is a collective behavior that consists of a large group of individuals moving together towards the same target direction. Unmanned Aerial Vehicle (UAV) flocking controllers have relied on Global Navigation Satellite System (GNSS) and intra-robot communication to obtain the absolute relative information of the nearby robots. This approach is only applicable in known and accessible outdoor environments. In this thesis, we explore the possibility of achieving flocking using a team of UAVs in hard-to-access locations, particularly with remote sensing restrictions. Thus, we propose a proximal control-based method for UAV self-organized flocking that relies on a vision-based relative localization approach proposed by Walter, Saska and Franchi (2018) called the Ultra-Violet Direction And Ranging (UVDAR) system. Robots use a Lennard-Jones potential function to maintain the cohesiveness of the flocking while avoiding collision within the teammates. After numerous simulations for safe verification and tuning, we evaluate our proposed method in a real-world environment with a group of middle-size UAVs using two distinct intra-swarm relative sensing approaches. In both cases, our method efficiently achieves flocking without alignment and direction control and moves into an arbitrary direction. In this way, we accomplished self-organized flocking with limited sensory information for aerial robots with high dynamics in environments with no constraints on the boundary conditions. As contributions, we have an extension of the work of Ferrante et al. (2012) and a decentralized flocking control method with local sensing capable of work in environments with GNSS restriction.