神经网络控制在综合火力／飞行系统中的应用

提出了两种基于模糊控制的神经网络控制器的设计方案，并将这两种控制器应用于综合火力／飞行系统的耦合控制。两种方法的主要差别在于获取样本的方式不同。方案１是通过对模糊控制方法得到的响应曲线采样获取样本，由Ｂａｃｋ－Ｐｒｏｐａｇａｔｉｏｎ学习算法训练神经网络，得到一组固定权值。神经网络控制器采用这组权值以“联想记忆”的方式工作。方案２则从用模糊控制算法得到的控制查询表中获取样本。因为模糊控制查询表比较大，采样时依据该表构成的相平面图的特点，对采样点数进行了压缩，使所设计的神经网络的规模可以接受，其余的设计步骤与方案１基本相同。仿真结果表明，采用这两种神经网络控制器的控制系统都具有良好控制性能

Application of the Neural Network Controller in the Integrated Fire/Flight Control System

Two design schemes of neural network controller based on fuzzy control are presented,which are applied to IFFCS (Integrated Fire and Flight Control System) coupling control.The fundamental difference of the two approaches lies in the way of obtaining samples. The first scheme is to get samplies through the response curves as a result of fuzzy control.The samples are used to train neural network by using the algorithm of Back Propagation. In consequence a group of definite weight values is obtained.These values are used by neural network controller for operating in the mode of association rememberance. The second scheme is to get samples from the control inquiry table which is produced by the fuzzy control algorithm. Because this table is too big, the sample points are compressed according to the property of phase plane of the table, so that the scale of designed neural network is acceptable.The other design steps of the two methods are almost similar. It is shown from simulation results that these two kinds of neural network controller have good control performance.