Identification of the Dynamic Parameters of a Parallel Kinematics Mechanism with Prismatic Joints by Considering Varying Friction

This study proposed a novel approach for the offline dynamic parameter identification of parallel kinematics mechanisms in which the friction is significant and varying. Since the friction is significant, it should be incorporated to provide an accurate dynamic model. Furthermore, the varying normal forces as a result of the changing posture of the mechanism lead to varying friction forces, specifically varying static and Coulomb friction forces. By considering this variation, the static and Coulomb friction parameters are identified as coefficients instead of forces. A bound-constrained optimization technique using an iterative global search tool was employed in this work to minimize the residual errors while maintaining the physical feasibility of the solutions. Moreover, the friction was modeled by using the nonlinear Stribeck friction model since a linear friction model was not sufficient, whereas the variation of the friction followed the variation of the normal forces, which were evaluated through the Lagrange multipliers in the constrained dynamic model of the mechanism. The solutions obtained were verified by using some trajectories that were different from those used in the identification.