Toronto Metropolitan University
Browse
Virtual_Decomposition_Based_Modeling_for_Multi-DOF_Manipulator_With_Flexible_Joint.pdf (9.62 MB)

Virtual Decomposition Based Modeling for Multi-DOF Manipulator With Flexible Joint

Download (9.62 MB)
journal contribution
posted on 2024-03-07, 20:12 authored by Kerui Xia, Hongjun Xing, Liang Ding, Haibo Gao, Guang Jun LiuGuang Jun Liu, Zongquan Deng

This paper discusses a problem that has plagued researchers for a long time regarding the dynamic modeling of a multiple degree-of-freedom (multi-DOF) manipulator such that its manipulation exhibits a higher computational efficiency and accuracy. Unknown friction, unknown gravitational torque, an uncertain moment of inertia, and severe joint coupling are the primary disturbing factors in multi-DOF manipulator modeling. In addition, joint flexible problems caused by the integration of harmonic drives increase the modeling complexity. Hitherto, no effective method has been found to address these problems. The virtual decomposition (VD)-based method exhibits the advantages of joint dynamics decoupling and minor computation compared with the traditional Lagrangian formulation or Newton–Euler formulation. In this study, an estimation method for the deformation-related torque of harmonic drives is established based on a novel experimental model; subsequently, this method is utilized in the VD-based model for the multi-DOF manipulator. Hence, the decoupling dynamic model for the manipulator considering joint flexibility is established. The performance of this new method has been evaluated by a contrast simulation with the Newton–Euler formulation, and the multi-DOF manipulator control simulation and experiment have been conducted with a VD-based model as a feedforward compensator to verify its performance in real-time control. The results demonstrated the validity and efficiency of e proposed approach. 

History

Language

English

Usage metrics

    Aerospace Engineering

    Licence

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC