神经网络

The work covers the control algorithm with neural network, the hardware realization of neural network and the design of distributed 3D animation real-time flight simulation system applied to attack helicopter.

内容涉及神经网络控制算法研究，神经网络的硬件实现及分布式武装直升机三维动画实时飞行仿真系统的设计等方面。

First, the recursive predictive errorlearning algorithm for recurrent neural networks is introduced, and its stability is demonstrated.

摘要针对递归神经网络BP学习算法收敛慢的缺陷，提出一种新的递归神经网络快速并行学习算法。

This dissertation highlights neural network models and learning algorithms related, which are playing an equally important role in neural network research. The main results obtained are presented as follows: Reproducing kernel developed in different disciplines has become an importanttool in functional approximation.

神经网络模型和与之相适应的训练算法是本文研究的重点，它们在神经网络的研究中占有同样重要的位置，本文获得的主要结果如下：再生核源于不同学科分支，目前已成为函数逼近的重要工具。

However, there are very few published studies dealing in neural network used in fault diagnosis of the locomotive and rolling stock.

论文在国内首次系统地将神经网络引入摆式列车倾摆控制系统的故障诊断中，从工程应用的角度深入研究了基于神经网络的故障诊断理论，探讨了其在摆式列车倾摆控制系统设计中的应用问题。

The experimental results show that the predicting result of RBFNN is better than that of BPNN. Root mean square error of prediction decreases from 0.72 0Brix to 0.65 0Brix.

研究结果表明：径向基函数神经网络比反向神经网络获得的预测结果更好，校正模型的预测均方根误差从0.720Brix降低到0.650Brix。

According to the input of the sample set, a function set of multiresolution wavelet is obtained and the output vectors of each node are also computed.

总结了多层前向神经网络的结构设计方法，然后提出了一种新的小波神经网络构造算法。

There are two ways to improve the robustness of the speaker recognition system. One is to select a proper type of neural networks, to optimize the parameters, so that it can maintain high performance of sample representation even if the features of voice signal was overlapped on sample space.

本文在对神经网络、话者识别，尤其是鲁棒话者识别的研究现状和发展进行了认真详细的分析的基础上，利用神经网络构建特定话者的特征映射器来改善话者识别系统的鲁棒性。

Moreover, online measurement system can be obtained after a fitting formula about water vapour saturated pressure in the gas pipeline at different temperatures is established by using the function chain neural network, which can greatly simplify the software of measurement of gas flux in the pipeline and quickly and truly realize measurement of gas flux in the pipeline in real time and on line in the design range of flowmeter.

在管道煤气计量系统测量中引入管道煤气相对湿度修正，并采用湿度传感器转换相对湿度信号，利用函数链神经网络对管道煤气工况温度下所对应的水蒸汽饱和压力进行拟合，得到基于函数链神经网络的管道煤气流量计量模型和在线计量系统，从而大大简化管道煤气流量计量软件，在流量计设计范围内实现管道煤气流量实时在线计量。

In Chapter 3 Firstly, the exponential stability in mean square and almost sure exponential stability for neutral stochstic neural networks with time-varying delay can be discussed by using the linear matrix inequality and the semimartingale convergence theorem; And our sufficient conditions are less conserative; Secondly, by utilizing the fixed point theorem, some sufficient and necessary conditions ensuring the asymp- totic stability in mean square for neutral stochastic neural networks with time-varying delays can be given.

第三章首先，利用线性矩阵不等式方法和半鞅收敛定理，考虑了一类具变时滞中立型随机神经网络的二阶矩指数稳定性和几乎必然指数稳定性，所得到的条件保守性要好；其次，采用不动点方法，给出了具变时滞中立型随机神经网络的二阶矩渐近稳定性的充要条件。

The sensitivity analysis is one of the popular research points of neural networks.

神经网络的敏感性分析一直是神经网络研究的热点之一。

We have no common name for a mime of Sophron or Xenarchus and a Socratic Conversation; and we should still be without one even if the imitation in the two instances were in trimeters or elegiacs or some other kind of verse--though it is the way with people to tack on 'poet' to the name of a metre, and talk of elegiac-poets and epic-poets, thinking that they call them poets not by reason of the imitative nature of their work, but indiscriminately by reason of the metre they write in.

索夫农 、森那库斯和苏格拉底式的对话采用的模仿没有一个公共的名称；三音步诗、挽歌体或其他类型的诗的模仿也没有——人们把&诗人&这一名词和格律名称结合到一起，称之为挽歌体诗人或者史诗诗人，他们被称为诗人，似乎只是因为遵守格律写作，而非他们作品的模仿本质。

The relationship between communicative competence and grammar teaching should be that of the ends and the means.

交际能力和语法的关系应该是目标与途径的关系。