Identification of a mechatronic motor-table system by using deep learning model

In this paper, system identification by the deep learning model (DLM) is proposed to compare with the self-learning particle swarm optimization (SLPSO) for a mechatronic motor-table system. Firstly, the complete dynamic formulation containing both mechanical equation with nonlinear frictional force and electrical equation is successfully formulated. Secondly, the governing equations are employed in the DLM and SLPSO to identify the unknown parameters for the mechatronic system. It is shown that the system identification can be successfully performed by using the DLM and SLPSO in this paper. In numerical simulations and experimental results, we discuss their advantage and disadvantage, and it is found that the DLM can identify the unknown parameters better converging toward the real ones when comparing with the SLPSO.