Neural fixed-time control framework for mobile robot platform via sliding mode method

Fixed-time control finds wide applications in various systems owing to its capability to achieve rapid convergence independent of initial conditions. This article investigates the fixed-time trajectory tracking control problem of the four-Mecanum-wheel mobile robot (FMWMR). First, taking into account the effects of parameter uncertainties and external perturbation, a second-order state-space model is derived to describe the FMWMR's dynamics and kinematics, based on which the tracking error system is built. Second, since the exact bound of the uncertain coupled vector which combines the parameter uncertainties and external perturbation is often not accessible in practice, the radial basis function neural network is employed to approximate it. To improve the FMWMR system's robustness and convergence rate, a non-singular fast terminal sliding-mode surface is synthesised for the trajectory tracking controller design. Sufficient conditions are derived to guarantee the practical fixed-time stability for the tracking error system. Finally, numerical simulations and physical experiments show the effectiveness and superiority of the proposed scheme and this work provides an initial-state-independent control framework that significantly advances the practical deployment of the FMWMR in complex industrial scenarios.