人工神经网络的单调序列逼近

设Nn,Ф是以Ф为激活函数的具有n+1个神经元的前向单隐层人工神经网络的全体.主要证明了,若f∈C[0,1],则对于任意的ε>0,存在两个神经网络序列{Pn,Ф}和{Qn,Ф},使得在[0,1]上Qn,Ф(x)≤Qn+1,Ф(x)≤f(x)≤Pn+1,Ф(x)≤Pn,Ф(x),而且Pn,Ф(x)-Qn,Ф(x)≤(6+4(2~(1/2)))En,Ф(f),这里的En,Ф(f)为Nn,Ф中的元对f的最佳逼近.     

基于神经网络的侧向激光雷达信号去噪算法

针对侧向激光雷达应用于气溶胶探测领域时,雷达回波信号易受噪声影响这一问题,本文提出了一种基于神经网络的激光雷达信号去噪算法。该算法在卷积神经网络基础上融合残差学习法和批量标准化,引入了注意力机制,改进激活函数,提升了网络性能和学习效率。采用本文提出的方法对噪声进行预测,实现了信号和噪声的有效分离,提高了侧向激光雷达CCD图像的信噪比。实验结果表明,使用本文提出的去噪算法对侧向激光雷达CCD图像进行去噪,图像的峰值信噪比提高了约5 dB,信号相对误差减小至9.62%,本文提出的去噪算法优于小波变换、维纳滤波等去噪方法,验证了该方法的可行性和实用性。

     

Electromagnetic NDE signal inversion by function-approximation neural networks

In the magnetic flux leakage (MFL) method of nondestructive testing commonly used to inspect ferromagnetic materials, a crucial problem is signal inversion, wherein the defect profiles must be recovered from measured signals. This paper proposes a neural-network-based inversion algorithm to solve the problem. Neural networks (radial-basis function and wavelet-basis function) are first trained to approximate the mapping from the signal to the defect space. The trained networks are then used iteratively in the algorithm to estimate the profile, given the measurement signal. The paper presents the results of applying the algorithm to simulated MFL data.     

Measurement of the lidar signal fluctuation with a shot-per-shot instrument

A measurement of the signal noise was carried out with a shot-per-shot lidar. This system was operated in the UV spectral region for ozone profiling in the low troposphere. We report on important discrepancies between our results and the estimations based on the assumptions commonly supporting the numerical modeling of lidar experiments.     

Coherent Doppler lidar signal spectrum with wind turbulence

The average signal spectrum (periodogram) for coherent Doppler lidar is calculated for a turbulent wind field. Simple approximations are compared with the exact calculation. The effects of random errors in the zero velocity reference, the effects of averaging spectral estimates by use of multiple lidar pulses, and the effects of the range dependence of the lidar signal power over the range gate are included. For high spatial resolution measurements the lidar signal power is concentrated around one spectral estimate (spectral bin), and correct interpretation of the contribution from turbulence is difficult because of the effects of spectral leakage. For range gates that are larger than the lidar pulse volume, the signal power is contained in many spectral bins and the effects of turbulence can be determined accurately for constant signal power over the range gate and for the far-field range dependence of the signal power.     

Numerical approximation of optimal control for distributed diffusion Hopfield neural networks

This work presents a numerical approximation of optimal control problems for non‐linear distributed Hopfield Neural Network equations with diffusion term. For one spatial dimensional case, a semi‐discrete numerical algorithm was constructed to find optimal control variable using finite element discretization, updated conjecture gradient iteration method. Furthermore, experiments demonstration will be implemented to show the effectiveness and stability through 3D graphics simulations. Copyright © 2006 John Wiley & Sons, Ltd.     

Prediction of thin film thickness of field emission using wavelet neural networks

Based on wavelet transforms extracting characteristic features from experimental data, the wavelet neural network (WNN) is used as an elementary model to study the characteristics of field emission from thin films. The WNN model is trained with learning samples of thin film thickness. The function mappings that the trained WNN model contains are the very ones that thin film thickness varies with characteristic parameters of field emission. A predicting model on thin film thickness of field emission is obtained. The data of thickness of diamond thin films is used to test this model. The results show that the absolute value of the relative error is within 2.98%, and the well-trained WNN model possesses good forecasting characteristics.     

A sequential approximation method using neural networks for engineering design optimization problems

There are three characteristics in engineering design optimization problems: (1) the design variables are often discrete physical quantities; (2) the constraint functions often cannot be expressed analytically in terms of design variables; (3) in many engineering design applications, critical constraints are often ‘pass–fail’, ‘0–1’ type binary constraints. This paper presents a sequential approximation method specifically for engineering optimization problems with the three characteristics. In this method a back-propagation neural network is trained to simulate a rough map of the feasible domain formed by the constraints using a few representative training data. A training data point consists of a discrete design point and whether this design point is feasible or infeasible. Function values of the constraints are not required. A search algorithm then searches for the optimal point in the feasible domain simulated by the neural network. This new design point is checked against the true constraints to see whether it is feasible, and is then added to the training set. The neural network is trained again with this added information, in the hope that the network will better simulate the boundary of the feasible domain of the true optimization problem. Then a further search is made for the optimal point in this new approximated feasible domain. This process continues in an iterative manner until the approximate model locates the same optimal point in consecutive iterations. A restart strategy is also employed so that the method may have a better chance to reach a global optimum. Design examples with large discrete design spaces and implicit constraints are solved to demonstrate the practicality of this method.     

Coherent Doppler lidar measurements of winds in the weak signal regime

In the weak signal regime coherent Doppler lidar velocity estimates are characterized by a localized distribution around the true mean velocity and a uniform distribution of random outliers over the velocity search space. The performance of velocity estimators is defined by the standard deviation of the good estimates around the true mean velocity and the fraction of random outliers. The quality of velocity estimates is improved with pulse accumulation. The performance of velocity estimates from two different coherent Doppler lidars in the weak signal regime is compared with the predictions of computer simulations for pulse accumulation from 1 to 100 pulses.     

基于小波包和PSO优化神经网络的刀具状态监测

为改进BP神经网络进行刀具磨损状态识别时训练时间较长、收敛速度较慢、测试精度不够高、甚至完全不能训练等问题,引入一种全局搜索能力强,收敛速度快的算法——粒子群优化算法(PSO),用其来优化BP神经网络参数,改进网络的训练和识别性能。实验证明:经粒子群算法优化后的BP神经网络较原网络有更快的训练迭代收敛速度和更高的测试准确度,达到优化的目标,对实现数控刀具磨损状态的智能化在线监测具有重要意义。     

基于神经网络和功率谱的水中目标信号检测方法

将功率谱和神经网络相结合，应用于高海况、低信噪比条件下，水中目标信号的特征提取中．文中首先对信号进行功率谱估计，利用目标信号功率主要集中在低频部分的特点，提取低频信号的能量作为特征，然后利用人工神经网络对目标信号进行检测．利用不同浪级情况下海洋水压场的仿真信号数据，对某型目标舰船的水压信号进行了检测计算，验证了该方法的有效性，尤其是达到了在高海况、低信噪比条件下，对目标信号检测率比较高、虚警率比较低的效果．     

基于小波变换的磨削噪声在线监测方法

随着自动化程度的提高,对磨削加工过程的在线监测方法研究越来越受到重视.无损检测技术的发展,提供了一种通过磨削噪声来在线监测砂轮磨削状况的方法.由于砂轮在磨削过程中产生的噪声,与其本身的材质和磨削状态有着密切的关系,因此通过对磨削噪声信号的分析,就可以精确地获取砂轮磨削状态的信息.将磨削加工过程中实时采集到的噪声信号进行...     

Experimental determination of the lidar overlap profile with Raman lidar

The range-dependent overlap between the laser beam and the receiver field of view of a lidar can be determined experimentally if a pure molecular backscatter signal is measured in addition to the usually observed elastic backscatter signal, which consists of a molecular component and a particle component. Two methods, the direct determination of the overlap profile and an iterative approach, are presented and applied to a lidar measurement. The measured overlap profile accounts for actual system alignment and for all system parameters that are not explicitly known, such as actual laser beam divergence and spatial intensity distribution of the laser light.     

基于小波奇异性的舰船辐射噪声音高分析研究

1引言 舰船辐射噪声包含着连续谱和线谱成分,一般说连续谱成分没有明确的音高仅对噪声的音色造成影响,但线谱成分在听测时人耳却可以感觉到明显的音高.声纳员在对舰船噪声听测时发现不同的舰船在音高值存在明显差别且音高的稳定性也有差异,因此通过舰船噪声音高值及稳定性对与舰船噪声识别非常有意义.     

Data analysis of stellar specklegrams with neural networks

An artificial neural network is applied to analysis of specklegrams of binary stars. Parameters of a binary star, the angular separation and the position angle, are estimated from the specklegrams by use of neural networks for each parameter. It is shown that a neural network is useful to analyze stellar specklegrams of binary stars.     

Radar signal design problem with neural network processing

Binary and ternary sequences with peaky autocorrelation, measured in terms of high discrimination and merit factor have been searched earlier, using optimization techniques. It is shown that the use of neural network processing of the return signal is much more advantageous. It opens up a new signal design problem, which is solved by an optimization technique called Hamming scan, for both binary and ternary sequences.     

Damage detection in aluminum and composite elements using neural networks for Lamb waves signal processing

One of the important factors in the structural health monitoring systems is the amount of data that need to be analysed in real time. This study investigated the use of artificially deteriorated signals of Lamb waves in training the novelty detection (ND) system for the early damage detection. In this system Auto-associative Neural Networks were trained using principal components calculated on the basis of experimentally measured signals. The specimens studied relate to two different materials commonly used in the aerospace industry, i.e. aluminium and glass fibre reinforced polymer. Lamb waves measured in these specimens are a good example that the ND algorithm works correctly in case of simple as well as complex signals. Furthermore, it was found that the designed ND system remained sensitive and robust even when it used raw signals with a relatively low sampling rate, on a fairly narrow time window and even noised signals.     

An analysis of the equilibria of neural networks with linear interconnections

In this paper, we analyse the equilibria of neural networks which consist of a set of sigmoid nonlinearities with linear interconnections,without assuming that the interconnections are symmetric or that there are no self-interactions. By eliminating these assumptions, we are able to study the effects of imperfect implementation on the behaviour of Hopfield networks. If one views the neural network as evolving on the openn-dimensional hypercubeH = (0, 1) ⁿ , we have the following conclusions as the neural characteristics become steeper and steeper: (i) There is at most one equilibrium in any compact subset ofH, and under mild assumptions this equilibrium is unstable. In fact, the dimension of the stable manifold of this equilibrium is the same as the number of eigenvalues of the interconnection matrix with negative real parts. (ii) There might be some equilibria in the faces ofH, and under mild conditions these are always unstable. Moreover, it is easy to compute the dimension of the stable manifold of each such equilibrium. (iii) A systematic procedure is given for determining which corners of the hypercubeH contain equilibria, and it is shown that all equilibria in the corners ofH are asymptotically stable.     

Optimum shape design of arch dams for earthquake loading using a fuzzy inference system and wavelet neural networks

An efficient methodology is proposed to find the optimum shape of arch dams considering fluid-structure interaction subject to earthquake loading. The earthquake load is considered by time variant ground acceleration applied in the upstream–downstream direction of the arch dam. The optimization is carried out by particle swarm optimization, employing real values of design variables. To reduce the computational cost of the optimization process, two strategies are adopted. In the first strategy, the most influential design variables on arch-dam response from original variables are selected using an adaptive neuro-fuzzy inference system. In the second, arch-dam response is predicted by a properly trained wavelet radial basis function neural network employing the influential design variables as the inputs. In order to assess the effectiveness of the suggested methodology, a real arch dam is considered as a test example. The numerical results demonstrate the computational advantages of the proposed methodology for the optimal design of arch dams.