Project I-RoCS: automatic navigation system for a cleaning robot

Abstract

Robotics is an interdisciplinary exciting and competitive eld, one of the largest research areas in the world today. The robotics market has encountered signi cant development over the recent years and is anticipated to grow tremendously over the near future. Robots have become part of our daily lives, particularly autonomous mobile robots. These robots are capable of moving and communicating with the world that surrounds them, thereby opening up new possibilities and making possible new types of human-robot interaction. In this area, decision-making capabilities are in constant evolution and they are constantly subject to sudden alterations, with the goal of approaching the behavior of robots to human behavior based on sensory data. This is a task of high complexity, but the advances of the technologies used have made them possible to produce better outcomes, thereby leading to the development of solutions that are increasingly closer to be used in real scenarios. Pre-programmed or reactive actions, generate the illusion of an intelligent robot, but for a robot to be considered genuinely intelligent, it must learn and adapt to the world in which it is inserted. In order to create such a robotic agent, it is important to develop solutions to the complex problems of autonomous navigation. This Thesis presents the study and development of an autonomous navigation system to be integrated into a cleaning robot prototype to be developed in the context of the i-RoCS project. This thesis presents an analysis of the most common algorithms for autonomous navigation, focusing on the three most important problems: mapping, localization and navigation. These steps are essential so that the robot can act autonomously in the environment where it will operate. Several algorithms have been explored, providing more emphasis on the Cartographer Google algorithm. A proposal of an autonomous navigation system was developed and tested both in simulation and using a real robot. The experimental results presented in this document using the educational Turtlebot robot demonstrate that it is possible for it to be able to navigate independently in human-occupied environments without recourse to external location systems. The implemented solution can be an important contribution to i-RoCS prototype.

Robotics is an interdisciplinary exciting and competitive eld, one of the largest research areas in the world today. The robotics market has encountered signi cant development over the recent years and is anticipated to grow tremendously over the near future. Robots have become part of our daily lives, particularly autonomous mobile robots. These robots are capable of moving and communicating with the world that surrounds them, thereby opening up new possibilities and making possible new types of human-robot interaction. In this area, decision-making capabilities are in constant evolution and they are constantly subject to sudden alterations, with the goal of approaching the behavior of robots to human behavior based on sensory data. This is a task of high complexity, but the advances of the technologies used have made them possible to produce better outcomes, thereby leading to the development of solutions that are increasingly closer to be used in real scenarios. Pre-programmed or reactive actions, generate the illusion of an intelligent robot, but for a robot to be considered genuinely intelligent, it must learn and adapt to the world in which it is inserted. In order to create such a robotic agent, it is important to develop solutions to the complex problems of autonomous navigation. This Thesis presents the study and development of an autonomous navigation system to be integrated into a cleaning robot prototype to be developed in the context of the i-RoCS project. This thesis presents an analysis of the most common algorithms for autonomous navigation, focusing on the three most important problems: mapping, localization and navigation. These steps are essential so that the robot can act autonomously in the environment where it will operate. Several algorithms have been explored, providing more emphasis on the Cartographer Google algorithm. A proposal of an autonomous navigation system was developed and tested both in simulation and using a real robot. The experimental results presented in this document using the educational Turtlebot robot demonstrate that it is possible for it to be able to navigate independently in human-occupied environments without recourse to external location systems. The implemented solution can be an important contribution to i-RoCS prototype.