Correction of DC Motor Control Signals by Machine Learning in 1D Overhead Crane Vibration Control Transport Experiments Using Vibration Manipulation Functions as Reference Input

Abstract

We perform vibration control transfer of a one-dimensional (1D) overhead crane under open-loop finite-time-settling (oFTS) control using vibration manipulation function (VMF) as reference input. A DC motor is used as an actuator of the overhead crane, and the DC motor is controlled by input signal that realizes trajectories designed along VMF. In the case of simple determination of the input signal from VMF, the experimental equipment does not move like the design trajectory due to the influence of friction. Therefore, we introduced correction terms using machine learning to modify the input signal. We have confirmed that the experimental equipment operates along the various designed trajectories with several pendulum. It was found that vibration damping transfer of a 1D overhead crane is possible under oFTS control with VMF as a reference input.