Thesis Type: Postgraduate
Institution Of The Thesis: Middle East Technical University, Graduate School of Natural and Applied Sciences, Turkey
Approval Date: 2019
Thesis Language: English
Student: HAMID MAJIDI BALANJI
Supervisor: Ali Emre TurgutAbstract:
Real-time pose estimation of Tool Center Point (TCP) of industrial robots is very important in industrial robotic applications. The TCP pose information is used in many industrial robotic tasks such as calibration and control tasks for on-line TCP pose corrections and improvement of the TCP pose accuracy and repeatability. Nowadays, in industry and robotic research laboratories, laser trackers, optical CMMs, stereo vision and photogrammetry techniques are applied for TCP pose estimation purposes; however, each of these has their own problems that affect their pose estimation performance and efficiency in industrial robots. This study presents a novel model-based pose estimation method based on computer vision and augmented reality markers. The proposed system is able to estimate the pose of the Tool Center Point (TCP) of the industrial robotic manipulators in point-topoint applications. An innovative Rhombicuboctahedron ArUco Mapper (RAM) is designed for pose estimation of the TCP. The performance analysis of the proposed system proved its repeatability and accuracy. The absolute positional measurement accuracy of the proposed system in the robot base frame is in the range of 0.12- 0.48mm and for orientational measurements in the robot base frame are in the rangevi of 0.003-0.012o. The pose measurement accuracy and repeatability in the camera base frame show better results ±50μm-500μm for the positional accuracy and ±0.009o- 0.05o for the orientational accuracy. The proposed method is considered as a costeffective, high accuracy and repeatable TCP pose estimator for industrial robotic applications.