基于递归神经网络的预测模糊控制

为提高控制信息与实时状态的适应性，改善模糊控制品质，用传统模糊控制策略，根据当前时刻误差和预测误差变化值，预测下一时刻的控制输出和系统在未来时刻的误差,用递归神经网络预报系统未来输出值的功能，采用双系统交替控制模式．系统中包含1个模糊控制器和1个递归神经网络，一个工作，另一个学习，使控制系统具有自适应性．仿真结果表明，与常规模糊控制相比，预测模糊控制使超调减小，调节时间缩短．

Predictive Fuzzy Control Based on Recurrent Neural Network

Recurrent neural network can forecast future output of plant, based on the function, it can be achieved to predict next sampling period change-of-error. Based on the cultent error and predictive change-of-error, predictive fuzzy control output of fuzzy logic controller could be obtained by general fuzzy logic control rules, which make control input more fit to system states, improve the control quality. Simultaneity, control model of double systems control in turn were adopted, each system included a FLC and a recurrent neural network, and ensured one system working the other learning at the same time all along, which makes control systems have property of adaptive. The computer simulation results indicate that predictive control improves dynamic quality of system.