Skip to search boxSkip to navigationSkip to main content

Lie group closed-form formulation of non-redundant dynamic model for generalized momentum based force estimation in branched articulated robots

  • Chao Tan(corresponding author)
    ,
  • Huan Zhao
    ,
  • Li Ding
    ,
  • ,
  • Kesheng Wang
    ,
  • Guohong Dai
*Corresponding author for this work
Research Output:
Contribution to journal
Article
Research
Peer-review

Abstract

<p>Accurate dynamic modelling and efficient force estimation are essential for advanced control of branched articulated robots. However, conventional identification methods often suffer from parameter redundancy and poor scalability to complex topologies, limiting their accuracy and computational efficiency. This paper presents a unified Lie group-based framework for non-redundant dynamic modelling and generalized momentum-based force estimation in branched articulated robots. By introducing motion decomposition and generalized inertia mapping, the proposed approach systematically eliminates parameter redundancy and constructs a non-singular dynamic model for branched articulated robots with revolute and prismatic joints. Analytical closed-form expressions for independent wrenches and dynamic derivatives are derived using Lie group operators, enabling efficient parameter identification and noise-resilient force estimation. Two generalized momentum observers are developed: one leveraging the analytical derivative of the mass matrix, and another employing a modified Coriolis matrix to ensure skew-symmetry, both avoiding the need for acceleration measurements. Simulation and experimental results on a 13-DoF dual-arm robot and a 6-DoF industrial manipulator demonstrate that the proposed method achieves improved computational efficiency and lower force estimation errors compared with conventional approaches. The framework facilitates accurate, efficient, and generalizable dynamic identification and sensorless force estimation for complex robotic systems, supporting advanced compliant control and physical interaction.</p>

Publication Information

Output type

Research Output:
Contribution to journal
Article
Research
Peer-review

Original language

English

Article number

103575

Journal (Volume, Issue Number)

Mechatronics (Volume 119)

Publication milestones

  • Accepted/In press - 25/06/2026
  • E-pub ahead of print - 03/07/2026
  • Published - 03/07/2026

Publication status

Published - 03/07/2026

ISSN

0957-4158

Publication IDs

  • Scopus: 105043554845