论文部分内容阅读
针对某电液伺服系统存在非线性及时变性,难以对其进行精确控制的问题,提出了神经网络自抗扰控制方案。该方案利用神经网络所具有的可以实现任意复杂映射关系的能力,将非线性误差反馈控制规律中的比例系数与微分增益作为单神经元自适应控制器的权系数,通过单神经元的自学习功能进行在线调节;同时利用RBF神经网络作为辨识器,以辨识被控对象的梯度信息。通过MATLAB计算机仿真结果证明,该控制方案使系统具有响应速度快、稳态精度高、鲁棒性强等优点,并能够有效的抑制外界扰动。
Aiming at the problem that a certain electro-hydraulic servo system has nonlinear time-varying degeneration and it is difficult to control it accurately, a neural network ADRC control scheme is proposed. By using the neural network’s ability to realize any complicated mapping relation, the scheme uses the proportional coefficient and differential gain in the nonlinear error feedback control law as the weight coefficient of the single-neuron adaptive controller. Through the single-neuron self-learning Function online adjustment; RBF neural network at the same time as a recognizer to identify the controlled object gradient information. The result of computer simulation of MATLAB shows that this control scheme has the advantages of fast response speed, high precision of steady state and strong robustness, and can effectively restrain the external disturbances.