Development of a Practical ICP Outlier Rejection Scheme for Graph-based SLAM Using a Laser Range Finder
| Autor: | Jun Hyeong Jo, Chang-bae Moon |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
0209 industrial biotechnology
Computer science business.industry Mechanical Engineering Graph based ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Mobile robot 02 engineering and technology Simultaneous localization and mapping Industrial and Manufacturing Engineering Graph 020303 mechanical engineering & transports 020901 industrial engineering & automation 0203 mechanical engineering Outlier Robot Reference map Computer vision Artificial intelligence Electrical and Electronic Engineering business |
| Zdroj: | International Journal of Precision Engineering and Manufacturing. 20:1735-1745 |
| ISSN: | 2005-4602 2234-7593 |
| DOI: | 10.1007/s12541-019-00175-0 |
| Popis: | The reference map for a given environment is one of the most essential elements for a mobile robot to automatically control itself and find a path. Simultaneous localization and mapping (SLAM) is a technology that aids the autonomous movement of a mobile robot in a given environment by simultaneously creating a map and performing localization. Graph-based SLAM is the state-of-the-art solution. However, graph-based SLAM faces certain problems when constructing graphs because of the front-end processing failure that occurs when using Iterative Closest Points (ICP). When the robot is in motion, a deviation between the scan images of the environment appears in every scan. Therefore, it is necessary to perform matching of the two images by using an outlier rejection algorithm for a robust ICP algorithm. In this paper, we propose an ICP outlier rejection scheme to compare to scan images and select matching points and reject mismatched parts. The experimental results show that the proposed scheme is reliable and robust in practical environments. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink