View-based mobile robot navigation : a genetic algorithmic approach
This thesis examines the simultaneous localization and mapping (SLAM) problem for mobile robot navigation. To obviate the dependency on successful feature extraction, we developed an efficient and flexible genetic algorithmic map representation for view-based SLAM approaches. It does not rely on fea...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2011
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/46711 |
| _version_ | 1826112808976121856 |
|---|---|
| author | Dong, Jun Feng |
| author2 | Wang Dan Wei |
| author_facet | Wang Dan Wei Dong, Jun Feng |
| author_sort | Dong, Jun Feng |
| collection | NTU |
| description | This thesis examines the simultaneous localization and mapping (SLAM) problem for mobile robot navigation. To obviate the dependency on successful feature extraction, we developed an efficient and flexible genetic algorithmic map representation for view-based SLAM approaches. It does not rely on features and it is especially appropriate in a 3D environment. With this map representation, an efficient view-based SLAM approach: the Rao-Blackwellized Genetic Algorithmic Filter (RBGAF) SLAM is developed. Such a SLAM approach does not rely on features and it is capable of integrating arbitrary sensor and motion models. Further more, the approach can be implemented on a graphical processing unit with the development of a highly efficient parallel computing structure for RBGAF-SLAM. This significantly improves the processing speed so that real-time processing can be achieved. A set of simulation and experiments are presented to demonstrate its effectiveness and efficiency. The results verify that our approach achieved 3D real time SLAM in urban environment. |
| first_indexed | 2024-10-01T03:13:10Z |
| format | Thesis |
| id | ntu-10356/46711 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T03:13:10Z |
| publishDate | 2011 |
| record_format | dspace |
| spelling | ntu-10356/467112023-07-04T16:54:52Z View-based mobile robot navigation : a genetic algorithmic approach Dong, Jun Feng Wang Dan Wei School of Electrical and Electronic Engineering DRNTU::Engineering::Computer science and engineering::Mathematics of computing::Probability and statistics DRNTU::Engineering::Industrial engineering::Automation This thesis examines the simultaneous localization and mapping (SLAM) problem for mobile robot navigation. To obviate the dependency on successful feature extraction, we developed an efficient and flexible genetic algorithmic map representation for view-based SLAM approaches. It does not rely on features and it is especially appropriate in a 3D environment. With this map representation, an efficient view-based SLAM approach: the Rao-Blackwellized Genetic Algorithmic Filter (RBGAF) SLAM is developed. Such a SLAM approach does not rely on features and it is capable of integrating arbitrary sensor and motion models. Further more, the approach can be implemented on a graphical processing unit with the development of a highly efficient parallel computing structure for RBGAF-SLAM. This significantly improves the processing speed so that real-time processing can be achieved. A set of simulation and experiments are presented to demonstrate its effectiveness and efficiency. The results verify that our approach achieved 3D real time SLAM in urban environment. DOCTOR OF PHILOSOPHY (EEE) 2011-12-23T06:01:25Z 2011-12-23T06:01:25Z 2011 2011 Thesis Dong, J. F. (2011). View-based mobile robot navigation : a genetic algorithmic approach. Doctoral thesis, Nanyang Technological University, Singapore. https://hdl.handle.net/10356/46711 10.32657/10356/46711 en 203 p. application/pdf |
| spellingShingle | DRNTU::Engineering::Computer science and engineering::Mathematics of computing::Probability and statistics DRNTU::Engineering::Industrial engineering::Automation Dong, Jun Feng View-based mobile robot navigation : a genetic algorithmic approach |
| title | View-based mobile robot navigation : a genetic algorithmic approach |
| title_full | View-based mobile robot navigation : a genetic algorithmic approach |
| title_fullStr | View-based mobile robot navigation : a genetic algorithmic approach |
| title_full_unstemmed | View-based mobile robot navigation : a genetic algorithmic approach |
| title_short | View-based mobile robot navigation : a genetic algorithmic approach |
| title_sort | view based mobile robot navigation a genetic algorithmic approach |
| topic | DRNTU::Engineering::Computer science and engineering::Mathematics of computing::Probability and statistics DRNTU::Engineering::Industrial engineering::Automation |
| url | https://hdl.handle.net/10356/46711 |
| work_keys_str_mv | AT dongjunfeng viewbasedmobilerobotnavigationageneticalgorithmicapproach |