Detalhes bibliográficos
| Ano de defesa: |
2018 |
| Autor(a) principal: |
OLIVEIRA, André Felipe da Silva
 |
| Orientador(a): |
ALMEIDA NETO, Areolino de
|
| Banca de defesa: |
ALMEIDA NETO, Areolino de
,
RIBEIRO, Paulo Rogério de Almeida
,
SANTOS, Sérgio Ronaldo Barros dos
,
ALMEIDA, Will Ribamar Mendes
|
| Tipo de documento: |
Dissertação
|
| Tipo de acesso: |
Acesso aberto |
| Idioma: |
por |
| Instituição de defesa: |
Universidade Federal do Maranhão
|
| Programa de Pós-Graduação: |
PROGRAMA DE PÓS-GRADUAÇÃO EM CIÊNCIA DA COMPUTAÇÃO/CCET
|
| Departamento: |
COORDENAÇÃO DO CURSO DE CIÊNCIAS DA COMPUTAÇÃO/CCET
|
| País: |
Brasil
|
| Palavras-chave em Português: |
|
| Palavras-chave em Inglês: |
|
| Área do conhecimento CNPq: |
|
| Link de acesso: |
https://tedebc.ufma.br/jspui/handle/tede/2424
|
Resumo: |
This work proposes a matching method of occupancy grid maps in image form, in which a new way of searching for similarities is developed.The maps used in this work were obtained via SLAM and were performed individually by different robots in unknown initial poses. The designed method is based on image processing techniques in which the maps characteristics are provided by robots and, through the relations between these characteristics, an alphabet is created to each map. The matching points candidates are found from the comparison between the members of the alphabet generated. After the comparison, it is possible to verify possible matches from the candidate points, where a similarity verification metric is applied. In this way, matching points that provide greater similarity are chosen as best current points. These operations are repeated every time new map updates are provided to the method. Thus, the similarity rates of the best points of the previous iteration are updated and new best matching are calculated for the current iteration, so that the matching points with the highest similarity rate between the previous and the current iteration are chosen as best current points. Several randomly generated environments with a varied amount of compartments were used to perform the tests. Several mappings were also performed to each environment. The tests considered the initial poses of unknown robots and the lack of communication between robots to share any information other than the current state of their maps. The results obtained are promising, since in the majority of the tests performed, it was successful in finding the correct matching between the maps. In some of these cases, the matching was found even in a situation of great deformation on the map, which is due to the skidding of the robots and noises in the measurements. |
