Application of subspace state-space identification methods on actuated multibody systems

In the last years subspace-based system identification methods have been developed for the application in open and even closed loop conditions. In this paper some of those methods are selected and applied to a nonlinear mechanical system in order to identify a local linear state space description of the underlying system. The considered plant is an actuated multibody system with two inputs and two outputs. The nonlinear nature of such a system leads to unstable behavior in some operating conditions so the system has to be operated in closed loop control during the identification procedure too. Considering these demands for the process of system identification the capability of subspace-based methods to identify stable and unstable MIMO systems using input-output data obtained from closed loop control experiments is investigated.