IEEE Trans Cybern. 2023 Apr;53(4):2636-2646. doi: 10.1109/TCYB.2022.3164739. Epub 2023 Mar 16.
In this article, a robust adaptive fixed-time sliding-mode control method is proposed for robotic systems with parameter uncertainties and input saturation. First, a model-based fixed-time controller is designed under the premise that the system parameters are known. Moreover, the unknown dynamics of robotic systems and the boundary of compounded disturbance are synthesized into a compounded uncertainty. Then, the Gaussian radial basis function neural networks (NNs) are selected to approximate the compounded uncertainty. In addition, the nonsingular fast terminal sliding-mode (NFTSM) control is incorporated into the proposed fixed-time control framework to enhance the robustness and convergence speed of unknown robotic systems. Finally, a comparative simulation based on a rigid manipulator shows the superiority and efficacy of the designed methods.
本文提出了一种针对具有参数不确定性和输入饱和的机器人系统的鲁棒自适应固定时间滑模控制方法。首先，在系统参数已知的前提下，设计了基于模型的固定时间控制器。此外，将机器人系统的未知动态和复合干扰的边界综合成一个复合不确定性。然后，选择高斯径向基函数神经网络 (NNs) 来逼近复合不确定性。此外，将非奇异快速终端滑模 (NFTSM) 控制纳入所提出的固定时间控制框架中，以增强未知机器人系统的鲁棒性和收敛速度。最后，基于刚性机械手的对比仿真表明了所设计方法的优越性和有效性。
