Development of a SLAM based Navigation System with Highway Implementation

• Main Authors: Jia-Lin Chen, 陳家霖
• Other Authors: Chyi-Yeu Lin
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/4w7n2x

碩士 === 國立臺灣科技大學 === 機械工程系 === 107 === The main purpose of this research is to develop a highway implementation to be used in navigation system that enables autonomous mobile robot to perform navigation task more quickly in indoor environments. In this research, the robot uses RGB-D SLAM as the main localization method. The image acquired by the RGB-D camera is used to obtain the detection feature points by the SURF algorithm and establish the 3D model of the work environment. This research develops a ROS-compatible program for fast and simple highway construction and implementation. When the robot's moving target is set, the highway system is used to accelerate the speed of indoor navigation and improve the stability to navigation movement. The accuracy of RGB-D SLAM often affects the stability of localization due to camera angle or environmental feature. The system design with a localization optimization method that can adapt to unstable localization, which can effectively and stably control the speed of the robot. In addition to specifying the location in the map, the navigation target can also be used with the object search system for instant object-searching navigation. The navigation system developed in this research, in addition to RGB-D SLAM localization methods, can also be combined with other positioning methods such as Lidar SLAM, RGB SLAM, and localization methods based on deep learning techniques.