Study of SLAM algorithms for autonomous navigation in unstructured environments

Robotics is one of the most exciting areas that has been through constant innovation and evolution over the years. Robots have become more and more a part of our lives and are no longer a vision for the future but a reality of the present. Nowadays we have robots cleaning our home, vacuuming our flo...

ver descrição completa

Detalhes bibliográficos
Autor principal: Ferrão, José Manuel Miranda (author)
Formato: masterThesis
Idioma:eng
Publicado em: 2019
Assuntos:
Texto completo:http://hdl.handle.net/10773/25126
País:Portugal
Oai:oai:ria.ua.pt:10773/25126
Descrição
Resumo:Robotics is one of the most exciting areas that has been through constant innovation and evolution over the years. Robots have become more and more a part of our lives and are no longer a vision for the future but a reality of the present. Nowadays we have robots cleaning our home, vacuuming our floors, playing soccer or even exploring the unknown outside of our planet. Robots are a major theme in research projects with special attention given to mobile robots since they have the capability to navigate the environment and interact more easily with humans. In the last couple of years, we have observed a big growth in the market of service robots. A service robot is dedicated to help humans in their everyday tasks. While reactive or pre-programmed behaviors are sufficient to let a robot appear intelligent, to be truly be intelligent a robot must learn and adapt to its environment. SLAM is the computational problem of learning an environment by constructing its map while simultaneously keeping track of the robot location inside it. Follow Inspiration is a company focused on the development of robotic systems. The autonomous shopping cart WiiGo was its first product, it is an autonomous service robot designed to help people carry their purchases in supermarkets. In this document we describe the testing and integration of SLAM algorithms, development of a marker-based solution to detect interest-points and the development of visualization tools for the WiiGo robot. The results presented in this document allowed the WiiGo robot to become capable of autonomous navigation in human occupied environments independently, without resourcing to external localization systems.