Kinematic sensitivity, parameter identification and calibration of a non-fully symmetric parallel Delta robot. (July 2021)

• Record Type: Journal Article
• Authors: Shen, Huiping; Meng, Qingmei; Li, Ju; Deng, Jiaming; Wu, Guanglei
• Is Part Of: Mechanism and machine theory. Volume 161(2021)
• Publication Date: 2021-07
• Subjects: Parallel Delta robot -- Sensitivity analysis -- Variational method -- Parameter identification -- Kinematic calibration; Machine theory -- Periodicals; Machinery -- Periodicals; Machines -- Périodiques; Génie mécanique -- Périodiques; Machine theory; Machinery; Periodicals; 621.81
• DOI: 10.1016/j.mechmachtheory.2021.104311
• Languages: English
• ISSNs: 0094-114X

Abstract

Highlights: Variational method to model sensitivity considering more error sources as possible. A modified Delta robot is less sensitive to the geometric variations than its original counterpart. Confirmed calibration method and parameter identification model by simulated calibration. Parameter identification by minimizing square residuals between predicted and measure positions. Robot absolute accuracy improved from 1.4mm to 0.1-0.13mm after calibration. Abstract: In this paper, the sensitivity analysis and kinematic calibration of a modified parallel Delta robot are presented. The influence of the variations related to the manipulator components and geometric parameters onto the motion accuracy, including the parallelogram, is studied by means of the variational method. The sensitivity analysis shows that the robot under study is slightly less sensitive to the geometric variations compared to its original counterpart. The developed error model was further used in the kinematic calibration of the robot as the parameter identification model, allowing more parameters and variations to be identified compared to the differential kinematic geometry based error models, which was confirmed through simulated calibrations. The calibration process was carried out along with a laser tracker to measure the end-effector positions, to identify the parameters by making use of the least square minimization method to minimize the square residuals between the predicted and measured positions. After calibration, the positioning error of the robot end-effector is significantly enhanced, with the absolute accuracy improved from 1.4 mm to 0.1– 0.13 mm . The presented calibration process can be applicable to other parallel robots.