Experimental identification of friction and dynamics of an overconstrained 3-DOF decoupled parallel mechanism

Obtaining a dynamic model for the parallel mechanism is always regarded as a challenging process due to high complexity of the dynamic equations especially in presence of friction. In this paper the process of black-box identification of friction and rigid-body dynamics of an overconstrained 3-DOF decouple mechanism is addressed. Since the under study mechanism is an overconstrained parallel mechanism, a theoretical model cannot be obtained. To this end, in order to avoid the problem of being overconstraint, i.e., inconsistent number of equations and unknown parameters in the dynamic model, the input/output black box identification technique is pursued. In this technique, an inverse dynamic model based on the 103 polynomial regressors of position, velocity, acceleration is proposed. Moreover, the friction of the each axes linear module is studied and modeled by a first order Fourier series and the friction terms are added to the 103 polynomial regressors model. Then, by resorting to the so-called Least Square method the latter model is identified. In addition, Forward selection is applied in order to obtain a simpler model with less number of regressors.