Study and implementation of a solution to the simultaneous location and mapping (SLAM) problem in mobile robots

• Grant number: 06/00968-1
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: May 01, 2006
• End date: December 31, 2006
• Field of knowledge: Engineering - Mechanical Engineering
• Principal Investigator: Jun Okamoto Junior
• Grantee: Vitor Campanholo Guizilini
• Host Institution: Escola Politécnica (EP). Universidade de São Paulo (USP). São Paulo , SP, Brazil

Abstract

The SLAM ("Simultaneous Localization and Mapping") is an attempt to solve one of the most fundamental problem of autonomous navigation nowadays, offering a way to approach it. In essence, SLAM tries to create a precise map of the surrounding environment of a mobile robot, even when great distances are covered. To achieve that, the knowledge of the actual location of the system at every update of the sensors is necessary, because the map is based on that information, and imprecise data will lead to imprecise maps. But, information relative to the localization are notoriously inaccurate, especially when obtained internally by the system, without the help from external equipment, which is precisely one of the cornerstones of autonomous navigation. This way, without a precise localization the results achieved by the mapping will also be imprecise, creating maps that do not correspond to the reality. The SLAM approach tries to accomplish both steps at the same time, the localization and the mapping, and thus allowing the creation of precise maps while securing the actual localization of the system at every step. This work has as final objective the implementation of an algorithm of SLAM, using its concepts and tools in order to allow an existent mobile robotic platform to navigate long distances while still maintaining a precise map of the environment it has already passed through, and at the same time being able to locate itself in this map. This solution will first be implemented in a virtual environment, using a dynamic model of the existing equipment, and then it will be validated in real tests, with the actual mobile robotic platform.