Mobile Robotics is an ever-expanding field in several application areas, but it is still
subject to numerous challenges, mainly related to localization in indoor environments.
Approaches commonly used to solve localization problems in these environments, such
as Simultaneous Localization and Mapping (SLAM), while quite effective in static environments,
are still easily subject to inaccuracies and failures in dynamic environments,
or environments with a very limited number of distinguishing features. In this work, an
approach that aims to integrate a SLAM technique based on particle filtering with a position
tracking algorithm for fiducial markers distributed in the indoor environment will
be presented. The proposed approach allows to perform corrections for intrinsic errors
accumulated in the particle filter, as well as a way to solve global localization problems
by obtaining the position of global landmarks based on the detection of the markers.
Experiments to validate the proposed localization system were performed in a simulation environment, which allowed to ensure the effectiveness of the method compared to the traditional approach with the particle filter. Based on the results obtained, the system was adapted for testing in a real environment, making use of a mobile robot integrated with the ROS framework.
