Neural Network Adaptive Hierarchical Sliding Mode Control for the Trajectory Tracking of a Tendon-Driven Manipulator

Abstract Tracking control of tendon-driven manipulators has become a prevalent research area. However, the existence of flexible elastic tendons generates substantial residual vibrations, resulting in difficulties for trajectory tracking control of the manipulator. This paper proposes the radial basis function neural network adaptive hierarchical sliding mode control (RBFNNA-HSMC) method, which combines the dynamic model of the elastic tendon-driven manipulator (ETDM) with radial basis neural network adaptive control and hierarchical sliding mode control technology. The aim is to achieve trajectory tracking control of ETDM even under conditions of model inaccuracy and disturbance. The Lyapunov stability theory demonstrates the stability of the proposed RBFNNA-HSM controller. In order to assess the effectiveness and adaptability of the proposed control method, simulations and experiments were performed on a two-DOF ETDM. The RBFNNA-HSM method shows superior tracking accuracy compared to traditional model-based HSM control. The experiment shows that the maximum tracking error for ETDM double-joint trajectory tracking is below 2.593×10-3 rad and 1.624×10-3 rad, respectively.