Experimental identification of the dynamics model for 6-DOF parallel manipulators

The paper presents a self-contained approach for the dynamics identification of six degrees of freedom (DOF) parallel robots. Major feature is the consequent consideration of structural properties of such machines to provide an experimentally adequate identification method. The known periodic excitation is modified and enhanced to take the actuator coupling as well as the numerical solution of the direct kinematics into account. The benefits of explicit frequency-domain data filtering are demonstrated. Additionally, a new implementation of the maximum-likelihood estimator allows for automatic tuning of the data filter. The issue of optimal input experiment design is also discussed and substantiated with extensive experiments.