RBFNN-based nonsingular fast terminal sliding mode control for robotic manipulators including actuator dynamics

Abstract To achieve robust finite-time trajectory tracking control, this paper proposes a novel neural-network-based nonsingular fast terminal sliding mode (NFTSM) control strategy for n-link robotic manipulators including actuator dynamics, subject to the model uncertainty and external disturbances. The suggested NFTSM control method can improve the finite-time convergence rate of system states, owing to the introduction of nonlinear item on the sliding surface. In addition, the singular problem is settled via introducing a saturation function into the control signal. In this control scheme, the precise dynamics of the robot system are unknown completely. Considering that the radial basis function neural network (RBFNN) has a fast study convergence speed and great approximation ability, three RBFNNs are utilized to estimate the manipulator-actuator dynamic parameters, along with an adaptive weight update law. Meanwhile, by designing robust control items, the approximation errors of RBFNNs are compensated, and the external disturbances are suppressed. Then, the finite-time stability of the controlled system is proved by Lyapunov stability theory. Finally, the proposed control approach is employed to a two-link robotic manipulator. The simulation results verified the effectiveness of the proposed control method.