Vision based control of a mobile robot

• Main Author: Zhang, Xiao
• Other Authors: Wang Jianliang
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/60513
• Institution: Nanyang Technological University

Vision-Based Robotic control is a revolutionary technology for new generation automation control. It uses the feedback information from the vision sensor to control the motion of a robot. The vision sensor used in the control system enables the robot with the ‘seeing’ ability to understand the unknown environment, which extend the range of usages for robots. Here in this project the Turtlebot2 is used as the mobile robot system with a Kinect camera attached on. The objective is to control the Turtlebot2 for certain motion perforce based by the vision information. Vision based tracking is one of the application which is implemented in the first part of this project. The OpenTLD is used for object tracking in a real-time video steam. Combine the TLD with a P controller, the Vision based object tracking control system is built up. Besides, the report also studied and implemented a non-vector space visual servoing method. It is a new way to conduct the vision based control. Non-vector space visual servoing does not require much image processing. It considers the image as a set with the elements of pixels holding the information of pixel location and intensities. In this method, the feedback error is defined the Hausdorff distance between the current-image set and desired-image set; and the control objective here is to minimize this feedback error. A set of new equations are then studied to solve this problem.