【摘要】 提出一种基于神经网络的模糊控制器．它可以把模糊控制的控制规则转化为多层前向神经网络的一对输入、输出样本．用Ｂａｃｋ－Ｐｒｏｐａｇａｔｉｏｎ学习算法对网络进行训练，使得网络记忆这些样本，并将这些样本以权值矩阵的形式存储在网络中．网络以“联想记忆”的形式来使用获得的经验对对象实施控制．知道了被控对象少量的定性知识，就可以用这种方法控制对象的行为．这种控制方案可用于对受控对象缺乏精确的数学描述或具有时滞、高阶等难以用现有的控制理论方法分析和控制的复杂系统．仿真结果证明了这种方法的有效性．更多还原

【Abstract】 A fuzzy controller based on neural network is proposed. It can transform the abstract experience rules of fuzzy logic control to a pair of input output samples of multilayer forward neural network.The network is first trained by Back Propagation algorithm and then made to memorize these samples which are stored in the network in the form of weight value. The controller utilizes these experiences to control the object according to the associative memory.By knowing the qualitative knowledge of the plant, its behavior can be controlled by this method. This controller makes the best of the superiority of the fuzzy logic control and neural network. The scheme can be applied to complicated systems ,which are difficult to control and to be analyzed by the control theory, such as absence of accuracy mathematical description , delay time or high order systems.The results of simulation have proved the scheme is efficient and feasible.更多还原