Simultaneous localization and mapping (SLAM) is one of the most important topics within the current robotics; two important processes for the development of SLAM; mapping and localization are defined in order to achieve the goal of building fully autonomous devices. This document focuses on the...

Simultaneous localization and mapping (SLAM) is one of the most important topics within the current robotics; two important processes for the development of SLAM; mapping and localization are defined in order to achieve the goal of building fully autonomous devices. This document focuses on the development of algorithms for processing point clouds to approach the SLAM concept in terms of mapping. Two algorithms for the registration process are proposed, which are based on two different operations; the first related to landmarks between each pair of samples to finally perform the calculation of the transformation matrix of point clouds, namely feature-based registration (FBR), the second based on the reduction of distances between points based on the ICP operation. Each algorithm achieves fairly good results regarding environment mapping, however the IPC based operation displays better results, as shown in the collected data

Fechado