Model-Free Control of Flexible Manipulator Based on Intrinsic Design

In this article, a model-free controller is proposed to suppress residual vibration caused by the linkage flexibility after the positioning process. The controller is designed to release the potential energy rapidly by repeating a sequence comprising two steps. In the first step, the controller acts as a damper to absorb the deformation energy, whereas in the second step, neural network control is introduced for the purpose of returning to the target position smoothly. The controller is simple, and the only required feedback signal is from the motor encoder. The position error is proved to converge to zero within a small neighborhood through Lyapunov stability analysis. The effectiveness of the controller is tested in three applications, an experiment study on a single-link flexible manipulator in horizontal plane and two simulation studies on a two-link flexible manipulator in horizontal plane and vertical plane, respectively. The settling time is decreased and energy efficiency is improved in all the three applications comparing with other controllers.