by Nidal Farhat, Miguel Diaz-Rodriguez, Vicente Mata
Abstract:
This paper deals with the identification problem of the inertial and frictional parameters of a parallel manipulator. Two important issue are considered; the physical feasibility of the identified inertial parameters and the use of linear and nonlinear friction models. The dynamic model, analyzed and reduced to a set of base parameters through the Singular Value Decomposition, is derived starting from the Gibbs-Appell equations of motion along with the Gauss principle of Least Action. The identification process is solved by using nonlinear constrained optimization techniques and it is verified, using exciting periodic trajectories, through a simulated parallel manipulator and applied over an actual 3-DOF RPS parallel manipulator. A comparison is made between the Least Square Method and the proposed optimization process in the case of linear friction models, and between the linear and nonlinear friction models in the optimization process.
Reference:
Dynamic parameter identification of parallel robots considering physical feasibility and nonlinear friction models (Nidal Farhat, Miguel Diaz-Rodriguez, Vicente Mata), In 12th IFToMM World Congress, 2007.
Bibtex Entry:
@inproceedings{farhat2007a,
title={Dynamic parameter identification of parallel robots considering physical feasibility and nonlinear friction models},
author={Farhat, Nidal and Diaz-Rodriguez, Miguel and Mata, Vicente},
booktitle={12th IFToMM World Congress},
address = {Besacon},
month = 6,
year={2007},
url={http://www.iftomm.org/iftomm/proceedings/proceedings_WorldCongress/WorldCongress07/articles/sessions/papers/A144.pdf},
gsid={https://scholar.google.com/scholar?oi=bibs&hl=es&cites=769629096827144864},
abstract={This paper deals with the identification problem of the inertial and frictional parameters of a parallel manipulator. Two important issue are considered; the physical feasibility of the identified inertial parameters and the use of linear and nonlinear friction models. The dynamic model, analyzed and reduced to a set of base parameters through the Singular Value Decomposition, is derived starting from the Gibbs-Appell equations of motion along with the Gauss principle of Least Action. The identification process is solved by using nonlinear constrained optimization techniques and it is verified, using exciting periodic trajectories, through a simulated parallel manipulator and applied over an actual 3-DOF RPS parallel manipulator. A comparison is made between the Least Square Method and the proposed optimization process in the case of linear friction models, and between the linear and nonlinear friction models in the optimization process.},
keywords={Nonlinear Friction model, Dynamic parameter identification, Optimization},
}