| Summary: | 碩士 === 國立雲林科技大學 === 資訊工程系 === 102 === ABSTRACT
In this thesis, an integrated 3D mapping approach based on RGB-D for a multi-robot system is implemented by ROS. This proposed system consists of three subsystems: Multi-robot Simultaneous Localization and Mapping Subsystem (MSLM), Integrated Map Construction Subsystem (IMC) and Human–computer interaction (HCI).
MSLM system is divided into three modules: (i) robot control module (ii) localization object and scene module (iii) mapping object and scene module. Robot control module is responsible for receiving user commands into robot motion control signals, so that the user can individually control two robots to move and SLAM. The MSLM subsystem localizes object and scenes by mapping object and scenes model and then extracts features information.
2D mapping is based on an occupancy grid method, Gmapping. Gmapping performs the Rao-Blackwellized particle filer and extended Kalman filter for robot localization. 3D mapping uses g2o optimizer to achieve globally consistent pose and output map model with optimized point cloud data.
In our implementation, the integrated 3-D map with RGB-D information of Multi-robot system is used in Department of School of Computer Science & Information Engineering, National Yunlin University of Science & Technology to construct the 2-D and 3-D map. The system not only provides an easy way for users to view 2-D map and 3-D map model of Department of School of Computer Science & Information Engineering but also provides a control interface to move the robots to specified location without any guiding.
Keywords: SLAM, localization, mapping, RGB-D sensor, kinect, multi-robot
|
|---|
