Scholarlink

<p>To provide safe boundary and precise control force for parallel mechanisms employed in large-scale motion simulation, the accurate analytic dynamic model is essential. However, the most challenge lies in the constitutive characterization of frictions with features of time-varying, nonlinearity, hysteresis, states dependence etc. To overcome the limitation of traditional empirical models, this study proposes a data-driven sparse identification method that regresses the analytic representation of frictions with both physical interpretability and precision. First, the dynamic model of the 6-PSU mechanism is established to calculate friction forces based on experimental data. Second, a friction model reconstruction method based on sparse model regression (SMR) incorporating iterative hyperparameter optimization is developed, successfully predicting the friction forces and exhibiting higher accuracy than the Coulomb-viscous (CV) model. After that, a full motion excitation trajectory experiment is carried out, with results showing the friction model’s extensionality for driving force compensation. Finally, based on the reconstructed friction model, the safety boundary for actuators is redefined. This study provides analytical model of a 6-PSU mechanism to ensure sufficiency of driving forces and safety from a dynamic perspective.</p> Graphical abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Friction force regression and driving force compensation of 6-PSU parallel mechanism based on data-driven sparse identification method

  • 期刊论文

  • 发表日期：

    2026年1月8日

查看全文
  • Zhiyao Yi,
  • Jiawei Qian,
  • Xiuting Sun,
  • Xiao Wang,
  • Xiaoxu Zhang,
  • Jian Xu

摘要

To provide safe boundary and precise control force for parallel mechanisms employed in large-scale motion simulation, the accurate analytic dynamic model is essential. However, the most challenge lies in the constitutive characterization of frictions with features of time-varying, nonlinearity, hysteresis, states dependence etc. To overcome the limitation of traditional empirical models, this study proposes a data-driven sparse identification method that regresses the analytic representation of frictions with both physical interpretability and precision. First, the dynamic model of the 6-PSU mechanism is established to calculate friction forces based on experimental data. Second, a friction model reconstruction method based on sparse model regression (SMR) incorporating iterative hyperparameter optimization is developed, successfully predicting the friction forces and exhibiting higher accuracy than the Coulomb-viscous (CV) model. After that, a full motion excitation trajectory experiment is carried out, with results showing the friction model’s extensionality for driving force compensation. Finally, based on the reconstructed friction model, the safety boundary for actuators is redefined. This study provides analytical model of a 6-PSU mechanism to ensure sufficiency of driving forces and safety from a dynamic perspective.

Graphical abstract