Vision-Aided Online Kinematic Calibration of a Planar 3R RR Manipulator
In this paper, we propose an automatic approach for the online kinematic calibration of a planar 3RRR manipulator. This method utilizes a camera as the global sensor to measure pose of the end effector. After the visual measuring system obtains enough pos
- PDF / 421,492 Bytes
- 10 Pages / 439.37 x 666.142 pts Page_size
- 29 Downloads / 169 Views
Abstract In this paper, we propose an automatic approach for the online kinematic calibration of a planar 3RRR manipulator. This method utilizes a camera as the global sensor to measure pose of the end effector. After the visual measuring system obtains enough pose data, an algorithm based on the implicit model is performed for the kinematic calibration. Finally, the calibration result is sent to the controller to compensate the geometric model of the manipulator. All these mentioned steps are implemented automatically without human intervene. Experiment studies on a real system have shown the effectiveness and the convenience of the present method.
Keywords Vision-aided Kinematic calibration 3RRR manipulator Automatic
1 Introduction Kinematic calibration is widely used in robotics to improve the absolute accuracy of manipulators. In other words, the basic goal of calibration is to find a more precise kinematic model of the manipulator. Usually, four steps are included in the calibration procedure [1, 2]. Firstly, developing a kinematic model. Secondly, measuring a sequence of poses with different configurations. Thirdly, identifying the accurate parameters of the models with proper optimized algorithm. Fourthly, compensating the controller with the optimized parameters. Due to the advantages of high stiffness, high speed, and compact structure, parallel mechanisms have drawn many researchers’ attention. Plenty of research about the kinematic calibration of parallel manipulators has been done in recent decades. Zhuang et al. present a method used for calibration of Gough-Stewart, measuring setup used in H. Li X. Zhang (&) L. Zeng H. Wu Guangdong Province Key Laboratory of Precision Equipment and Manufacturing Technology, South China University of Technology, Guangzhou 510640, China e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2017 X. Zhang et al. (eds.), Mechanism and Machine Science, Lecture Notes in Electrical Engineering 408, DOI 10.1007/978-981-10-2875-5_79
963
964
H. Li et al.
this method is theodolite [3]. Joubair et al. [4, 5] carry out calibrations for a planar PM and a five bar mechanism with the measuring setup of coordinate measurement machine (CMM) and measurement arm (MA). Kim presents a calibration for a Cartesian PM using a ballbar [6]. Except commercial measurement setups mentioned above, vision-based equipment also has widely used in this area. Renaud et al. [7] propose a calibration method for a H4 PM with a vision-based metrology. Daney et al. [8] propose an optimized pose choosing strategy and realize the calibration for a Gough-Stewart PM with a vision-based device. Traslosheros et al. [9] realize the calibration of a delta PM with a camera attached at the end effector. An obvious drawback for the calibration methods mentioned above is that this procedure is done before the use of the manipulator, which may cause the calibration result inaccurate when the manipulators are loaded or continuously running. In addition, as the measurement setup and the manipul
Data Loading...
