Banca de QUALIFICAÇÃO: LEONARDO ANGELO VIRGINIO DE SOUTO

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
DISCENTE : LEONARDO ANGELO VIRGINIO DE SOUTO
DATA : 06/07/2017
HORA: 09:00
LOCAL: Mini-auditorio do PPgEEC
TÍTULO:

Solution for Robotic Localization as a Quadratic Eigenvalue Problem Using the Concept of Natural Landmark Visual Memory.

PALAVRAS-CHAVES:

Recognition of Natural Frames, Mobile Robot, Vision Localization, Robotic Vision, Visual Odometry, Polynomial Problem of High-values

PÁGINAS: 40
GRANDE ÁREA: Ciências Exatas e da Terra
ÁREA: Ciência da Computação
SUBÁREA: Metodologia e Técnicas da Computação
ESPECIALIDADE: Processamento Gráfico (Graphics)
RESUMO:

With the progress in the research of mobile robots localization algorithms, the need for identification and recognition of natural reference points, as well as the importance os having better accuracy and computational performance have increased. Detection of natural landmarks is a challenging task because their appearance may be different in shape and design. In addition, they have mishaps in their identification, such as ambient lighting influence. The intent of this research becomes even more challenging, since the goal will be to use natural landmarks as elements of a visual memory geared to the location of any mobile robotic platform. Thus, this research will seek to innovate the knowledge and results in the field of robotic location, as well as to use applied linear algebra to obtain a better computational use. From these fields, we call the approach as a robotic visual memory algorithm for localization based on the detection, identification and correlational storage of natural landmarks of mobile robots having as input images acquired from a RGB-D camera. Our approach will simulate the way humans locate themselves, that is, the robot, when it arrives in what it finds as a new location, will keep visual characteristics in its memory that facilitate and minimize the errors of location and recognition of the place when it revisits that area. As a result, its ability to plan routes also has many possibilities, as you can stipulate the best and fastest route through the milestones you are currently viewing.

MEMBROS DA BANCA:
Presidente - 1345674 - LUIZ MARCOS GARCIA GONCALVES
Externo ao Programa - 1170845 - BRUNO MARQUES FERREIRA DA SILVA
Externo à Instituição - ANDERSON ABNER DE SANTANA SOUZA - UERN
Externo à Instituição - TIAGO PEREIRA DO NASCIMENTO - UFPB