Broad Learning System-Based Fixed-time Control for Robotic Manipulators with Input Deadzone

In this paper, a neural network trajectory tracking controller is proposed for robots with input deadzone. An improved radial basis function neural networks (RBFNNs) with broad learning system (BLS) is utilized for estimating the uncertain model of the system and the input deadzone function. By constructing Lyapunov functions, a fixed-time control scheme has been designed to ensure that the convergence time of the system is not affected by the initial state. The efficacy of the control method was verified by simulations. Through simulation, it is indicated that the proposed control method can resolve the problem of input deadzone and exhibit excellent tracking performance.