pp. 1013-1021
S&M1263 Research Paper of Special Issue https://doi.org/10.18494/SAM.2016.1390 Published: September 21, 2016 Data Association of Aerial Robot Monocular Simultaneous Localization and Mapping [PDF] Yin-Tien Wang, Chung-Hsun Sun, Ting-Wei Chen, and Chen-Tung Chi (Received December 24, 2015; Accepted July 6, 2016) Keywords: visual simultaneous localization and mapping (vSLAM), image depth parameterization, dataassociation, map management
This paper presents an algorithm for data association for the visual navigation of aerial robots.The major objective is to provide the aerial robot with the capabilities of localization and mappingin global positioning system (GPS) denied environments. The visual sensor system could measureinformation for robot state estimation and environmental mapping as the aerial robot navigates in aGPS-denied environment. Only one single camera was used to reduce the load on the aerial robot.The captured image was transmitted to a personal computer for image processing using a radiofrequency transmitter. In this study, an efficient data association method based on fuzzy rules wasdeveloped to determine the robust landmarks for robot mapping. An ultrasonic sensor was designedto provide distance measurements and to solve the map scale determination problem of monocularvision. The software program of the robot navigation system was developed on a windows-basedpersonal computer. The navigation system integrated the visual sensor, the algorithm for dataassociation, and the state estimator. The integrated system was used to carry out simultaneouslocalization and mapping for aerial robots.
Corresponding author: Yin-Tien WangCite this article Yin-Tien Wang, Chung-Hsun Sun, Ting-Wei Chen, and Chen-Tung Chi, Data Association of Aerial Robot Monocular Simultaneous Localization and Mapping, Sens. Mater., Vol. 28, No. 9, 2016, p. 1013-1021. |