论文部分内容阅读
本文提出了一种基于神经网络的ＰＩＤ控制器。利用递推最小二乘法在线整定ＰＩＤ参数，以克服神经网络ＢＰ算法收敛速度慢和可能出现的局部最小。仿真研究表明，这种ＰＩＤ控制器参数整定方便，控制精度高，跟随特性好，抗干扰能力强。
This paper presents a PID controller based on neural network. The recursive least square method is used to set the PID parameters online to overcome the slow convergence speed and the possible local minimum of the BP neural network algorithm. The simulation results show that this kind of PID controller has the advantages of easy parameter setting, high control precision, good following characteristics and strong anti-interference ability.