A novel nonlinear disturbance predictor based on reconstructed attractor for motion control system

We propose a novel nonlinear disturbance predictor to be used for motion control system. As an example, we develop a nonlinear disturbance predictor to compensate for the interference force of multiple axes robot manipulator. The reconstructed attractor which is often used in chaos analysis is applied to express and to predict disturbance. Using this predictor, high performance servosystem is realized without the exact plant model of robot dynamics. It is effective when there are unknown dynamics such as friction force or even chaotic behavior difficult to be modeled. Some simulation results using the double pendulum system are shown to verify the effectiveness of the proposed method