Hybrid neural network-based fractional-order sliding mode controller for tracking control problem of reconfigurable robot manipulators using fast terminal type switching law

A hybrid neural network-based fractional-order sliding mode controller for the position/force tracking control problem of a reconfigurable robot manipulator system is presented in this work. Due to interchangeable link modules, modeling uncertainties, coupled interconnected states, etc., the control of reconfigurable robot manipulators is very complicated in variable circumstances. So, to handle these dynamical systems, initially, a stable fractional-order sliding manifold is introduced to facilitate accurate and faster system state responses. Subsequently, a neural network-based fractional-order fast terminal sliding mode controller is designed to manage the consequences of external disruptions and parametric uncertainties effectively. In the controller’s design, a hybrid combination of radial basis function neural network and adaptive compensator with fast terminal type switching law is opted for robust performance of the dynamical system. The novelty of the work lies in the combination of the hybrid intelligent sliding mode control scheme with fractional calculus for tracking control problems of reconfigurable robot manipulator systems. The proposed scheme improves the transient response of the controller with a fast terminal-type switching law, and addresses the robustness, fixed-time convergence of system states along with an explicit assessment of the settling time. Finally, the asymptotic stability of the closed-loop dynamical system is validated through Lyapunov’s stability criteria and Fractional-order Barbalat’s lemma, and the simulation results along with a comparative study with some quantitative statistical evaluations confirm the contribution of the presented work.