Radial Basis Function Neural Network Based Fractional Order Sliding Mode Control for Planar Flexible-Joint Robot 2-Dof

In this paper, the authors propose a combination of a sliding controller using fractional order derivatives combined with neural networks to control a Flexible-Joint Robot system with two degrees of freedom under the influence of disturbances. In which disturbances are defined including the error of the robot's model parameters and the external disturbance. In addition, Matlab/Simscape Multibody combined with Solidworks is used to test the stability of a nonlinear planar Flexible-Joint Robot system with two degrees of freedom. By using a semi-physical model, the simulation evaluation is more accurate than the traditional mathematical model. The results show the stability of the proposed controller against the influence of uncertainties and external disturbance.