摘要神经网络是由大量人工神经元（处理单元）广泛互联而成的网络，它具有很强的自适应性和学习能力、非线性映射能力、鲁棒性和容错能力。随着被控系统变的越来越复杂，人们对于控制系统的要求也变的越来越高，需要控制系统能适应不确定性、时变的对象与环境下工作。传统的基于精确模型的控制方法难以适应要求，现在有关控制的概念也已更加广泛，它要求包括一些决策、规划以及学习功能。神经网络由于具有上述优点而越来越受到人们的重视。神经网络控制系统是一种把人工神经网络作为控制器或者辨识器的控制系统。神经网络在实际建模和控制领域中可以用作基于模型的各种控制结构，也可以本身用作控制器，同时也能在控制系统中起优化计算的作用。20800
本论文首先综述了神经网络的起源和发展，以及以神经网络作为控制器的典型神经网络控制模型。在此基础上，系统的研究了传统控制模型和神经网络控制模型频率特性的异同,因而得到神经网络控制系统的响应的快速性和平稳性都比传统控制系统要好。
关键词 神经网络 控制器 频率特性
毕业论文设计说明书（论文）外文摘要
Title The frequency characteristic of neural network control model
Abstract
Neural network has a strong adaptability and learning ability and nonlinear mapping ability, robustness and fault -tolerance,which is made up of a lot of artificial neuron.With the controlled system is more and more complex, people more and more high to the requirement of control system, especially the control system can adapt to uncertainty, time-varying objects and the environment.The traditional control system based on a accurate model is difficult to adapt to the requirements, now about the concept of control has more widely, it demands include some decision-making, planning and learning function.Neural network is more and more get people's attention.Neural network control system which makes artificial neural network as a controller or identifier, can be used for various control structure based on the model in the field of practical modeling and control, also can be used as the controller itself, as well as in control system plays the role of optimization calculation.
This paper first summarizes the origin and development of neural network, and the typical neural network control model which makes the neural network as a controller.On this basis, we studied the similarities and differences of frequency characteristic between the traditional control model and neural network control model , so we get that neural network control system is better than traditional control system in the response rapidity and stability.
Keywords Neural network controller frequency characteristic
目 次
1 引言 1
2 在系统建模和控制中常用的神经网络 3
2.1 BP神经网络 3
2.2 RBF神经网络 4
2.3 模糊神经网络 6
3 神经网络作为控制器的情况下常见的控制模型 8
3.1 模型参考自适应控制 8
3.2 前馈直接控制 9
3.3 损失函数型控制 9
3.4 估值器-控制器方案 10
4 基于BP神经网络的PID控制器 11
4.1 神经网络PID控制器的结构 11
4.2 神经网络PID控制器的控制算法 11
5 MATLAB中S函数的设计与应用 15 神经网络控制模型设计与仿真研究:http://www.youerw.com/zidonghua/lunwen_12854.html