提出了一种新的具有预设性能的自回归小波神经网络（self-recurrent wavelet neural network，SRWNN）超扭曲非奇异快速终端滑模（super-twisting non-singular fast terminal sliding mode，STNFTSM）控制方法（SRWNN_STNFTSM），在动力学不确定性和未知扰动的情况下提高板球系统的跟踪控制性能。利用预设性能函数（prescribed performance function，PPF），将板球系统受约束的位置误差转换为无约束的误差模型。引入非奇异快速终端滑模（non-singular fast terminal sliding mode， NFTSM）面来消除常规终端滑模控制存在的奇异问题，并加入一个tanh函数的补偿项改进NFTSM滑模面，以调节轨迹跟踪的收敛速度和跟踪精度，同时结合超扭曲算法（super-twisting algorithm，STA）设计STNFTSM控制器，以削弱抖振和集总扰动的影响。针对系统存在的集总扰动，为了保证高跟踪精度，结合STNFTSM设计了自适应SRWNN补偿器来消除扰动，保证了鲁棒性。与现有常规滑模控制相比，仿真验证表明SRWNN_STNFTSM具有良好的跟踪性能和鲁棒性，能够对集总扰动下的板球系统进行准确跟踪。

A novel control method, the self-recurrent wavelet neural network super-twisting non-singular fast terminal sliding mode (SRWNN_STNFTSM) control with prescribed performance, is proposed to improve the tracking control performance of the ball and plate system in the presence of dynamic uncertainties and unknown perturbations. The prescribed performance function (PPF) is used to convert the originally constrained position error of the ball and plate system into an unconstrained error model. The non-singular fast terminal sliding mode control (NFTSMC) sliding mode surface is introduced to resolve the singular issue of conventional terminal sliding mode control. Additionally, a compensation term of the tanh function is incorporated to improve the NFTSM sliding mode surface, adjusting the convergence speed and tracking accuracy. Moreover, the SRWNN_ STNFTSM controller is combined with the super-distortion algorithm (STA) to mitigate the effects of chattering and lumped disturbance. To address the lumped disturbance of the system and ensure high tracking accuracy, an adaptive SRWNN compensator is designed in conjunction with the STNFTSM. This compensator is aimed at eliminating disturbances and ensuring robustness. Simulation results compared with existing conventional sliding mode control methods demonstrate that SRWNN_STNFTSM exhibits excellent performance. It accurately tracks the ball and plate system under the influence of lumped disturbances.

TP273

国家自然科学基金资助项目(61163051)。