In this paper, the finite-time position tracking control for the fractional-order permanent magnet synchronous motor with input saturation, load disturbance and parameter uncertainties is considered. Based on the command filtered backstepping method, a novel adaptive neural network finite-time controller is presented. First, the neural network is introduced to approximate the uncertain function. Then, by using a command filter at the output side of the virtual signal, the issue of “explosion of complexity” is avoided. In addition, an adaptive law is applied to eliminate the approximation error and filtering error. Meanwhile, by utilizing the terminal sliding mode control technique, the finite-time signal tracking is achieved. Furthermore, an auxiliary design system is constructed to cope with the input saturation constraint. The proposed scheme not only possesses the superiorities of the classical command filtered backstepping, but also holds quick finite-time convergence property. Numerical simulations and experiment results are included to reveal the validity of the proposed design.

本文考虑了具有输入饱和，负载扰动和参数不确定性的分数阶永磁同步电动机的有限时间位置跟踪控制。基于命令滤波反推方法，提出了一种新型的自适应神经网络有限时间控制器。首先，引入神经网络来近似不确定函数。然后，通过在虚拟信号的输出侧使用命令滤波器，避免了“复杂性爆炸”的问题。另外，应用自适应定律来消除逼近误差和滤波误差。同时，通过利用终端滑模控制技术，实现了限时信号跟踪。此外，构造了辅助设计系统以应对输入饱和约束。所提方案不仅具有经典命令滤波反推的优势，而且具有快速的有限时间收敛性。数值模拟和实验结果包括在内，以揭示所提出设计的有效性。