基于模糊变步长神经网络的自适应PID控制

本文针对具有非线性和不确定性的控制对象，提出了一种基于前向神经网络的自适应PID控制系统算法。并根据其特点，提出学习步长模糊校正的方案.这种控制系统综合了神经网络、模糊决策和PID控制的优点，易于实现，仿真结果表明该控制系统对非线性被控对象具有比较好的控制效果。

An Adaptive FID Controller Baesd on Neural Network with Learning Rate Tuned by Fuzzy Rules

An adaptive PID Controller based on feed forward neural network is proposedin this paper.The PID parameters are tuned on line by the neural network and an auxiliarp controlsignal Produced by the neural network is added to the plant besides PID control signal.Thelearning rate of the neural network is tuned by fuzzy logic rules .The method was found to providesatisfactory control of noulinearsestem with strutural uncertainty.