基于新型PID神经网络的自适应控制系统研究

提出一种新的PID型神经网络的自适应控制系统,该控制系统采用对角递归神经网络辨识对象的正向模型,采用一种新型神经网络控制器产生控制量,与常规PID控制不同的是,该控制量不再是误差信号的比例、积分和微分量的简单线性组合,而是这些信号的一种非线性组合,从而可以有效地解决常规PID控制器存在的快速性和超调量之间的矛盾.仿真实验表明,这种新型控制系统具有较强的自适应性和鲁棒性.

Adaptive Control System Based on a New PID Neural Network

A new adaptive control system based on a new PID neural network is presented.The control system uses diagonal recurrent neural network to identify the forward model of the plant.The control quantity is produced by a new neural network controller.The difference from the conventional PID is that the control quantity is not the simple linear combination of the proportion,integral and differential coefficient of the error signal,but a nonlinear combination of them.So it solves the problem exiting in conventional PID control.The simulation result shows the control system has strong adaptive ability and robustness.