基于Chebyshev正交多项式的通道相位估计算法

针对信号传输通道相位估计等问题,提出了一种通过利用Chebyshev正交多项式逼近函数计算反正切函数值求取通道相位的算法,该算法通过象限判别、区间划分的方法计算反正切函数的输入值,然后利用Chebyshev正交多项式逼近函数计算信号相位,最后,通过迭代算法简化计算。该算法具有算法简单,估计速度快、精度高的特点。     

用于电磁兼容预测的函数链神经网络

为了解决传统BP神经网络的电磁兼容性预测方法存在易于局部收敛的问题,提出了一种基于模糊测度的函数链神经网络。该网络通过函数链将初始权重扩展到更高维度上,在实现传统BP网络多层感知的功能同时,计算过程仅为单层运算,因以收敛速度比多层的BP神经网络更快,解决了网络在解决非线性问题时,收敛于局部最小的问题。实验结果表明,提出的基于模糊测度的函数链神经网络在预测电磁兼容性方面更加精确。     

切比雪夫函数型连接神经网络在信道均衡中的应用

本文提出一种基于切比雪夫函数型连接神经网络(CFLNN)的信道均衡方法。传统的前馈神经网络虽然能有效地解决信道均衡的问题,但具有计算复杂度过高,收敛速度慢等缺点。函数型连接神经网络通过对输入模式进行非线性扩展,可以不必使用隐层而不降低整体性能,从而极大简化了网络结构。同时,神经网络的学习方法得以简化,提高了收敛速度。本文采用可变尺度共扼梯度下降法(SCG)对该函数型连接网络进行训练。仿真结果表明了用切比雪夫函数型连接神经网络解决信道均衡问题的有效性。     

AN ALGORITHM FOR VLSI CHANNEL ROUTING USING NEURAL NETWORK

The algorithm for VLSI channel routing using Hopfield neural model is discussed inthis paper.The basic methods of mapping VLSI channel routing problem to Hopfield neural net-work,constructing energy function,setting initial neural status,and selecting various parametersare proposed.Finally,some experimental results are given.     

小生境进化RBF网络在信道均衡中的应用

基于RBF网络的信道均衡器,其性能取决于神经网络的隐层参数.文中采用判决反馈均衡器的结构,RBF网络隐层的结构和参数则由基于小生境技术的混合递阶遗传算法来确定.仿真结果表明,这种RBF网络均衡器性能优良.     

Identification of DNS covert channel based on improved convolutional neural network

In order to effectively identify the multiple types of DNS covert channels,the implementation of different sorts of DNS covert channel software was studied,and a detection based on the improved convolutional neural network was proposed.The experimental results,grounded upon the campus network traffic,show that the detection can identify twenty-two kinds of data interaction modes of DNS covert channels and is able to identify the unknown DNS covert channel traffic.The proposed method outperforms the existing methods.     

（英）CCNet:面向多光谱图像的高速船只检测级联卷积神经网络

针对实现遥感图像中船只目标的快速检测,提出了一个采用多光谱图像、基于级联的卷积神经网络(CNN)船只检测方法CCNet。该方法所采用两级级联的CNN依次实现感兴趣区域(ROI)的快速搜索、基于感兴趣区域的船只目标定位和分割。同时,采用含有更多细节信息的多光谱图像作为CCNet的输入,能够提升网络提取特征鲁棒性,从而使得检测更加精确。基于SPOT 6卫星多光谱图像的实验表明：与当前主流的深度学习船只检测方法相比,该方法能够在实现高检测精准度的基础上将检测速度提高5倍以上。     

有线电视双向HFC网络上行信道噪声分析

以相关数学公式以及物理定律为依据，对有线电视双向HFC网络上行道噪声模型进行了理论推导，提出了解决噪声干扰的办法。     

Application of neural networks with orthogonal activation functions in control of dynamical systems

In this article, we present a new method for the synthesis of almost and quasi-orthogonal polynomials of arbitrary order. Filters designed on the bases of these functions are generators of generalised quasi-orthogonal signals for which we derived and presented necessary mathematical background. Based on theoretical results, we designed and practically implemented generalised first-order (k = 1) quasi-orthogonal filter and proved its quasi-orthogonality via performed experiments. Designed filters can be applied in many scientific areas. In this article, generated functions were successfully implemented in Nonlinear Auto Regressive eXogenous (NARX) neural network as activation functions. One practical application of the designed orthogonal neural network is demonstrated through the example of control of the complex technical non-linear system – laboratory magnetic levitation system. Obtained results were compared with neural networks with standard activation functions and orthogonal functions of trigonometric shape. The proposed network demonstrated superiority over existing solutions in the sense of system performances.     

基于GAN指纹库的卷积神经网络室内可见光信道模型

提出一种神经网络算法实现室内可见光信道模型,解决Lambert模型难以计算室内可见光信道的噪声和误差问题。针对指纹库数据量大、难以采集和训练参数多导致迭代速度慢的问题,提出使用生成式对抗网络(generative adversarial network,GAN)生成仿真数据集融合原有的稀疏指纹库,生成满足训练要求数量的指纹库;使用一维的卷积神经网络(convolutional neural network, CNN)提取数据特征,降低训练参数,提高迭代速度。在室内5 m×5 m×3 m环境下采集稀疏指纹库,分别用反向传播 神经网络(back propagation netural network, BPNN)和一维CNN室内可见光信道模型进行对比。仿真结果表明:使用GAN生成指纹库的平均绝对误差为0.04,对数据量增广300%;在同一指纹库下,BPNN信道模型误差为3.81,迭代500次收敛;而CNN信道模型误差为0.79,迭代100次收敛。本文提出的GAN指纹库融合CNN的可见光信道模型具有精度高、误差小、速度快、泛化性强等优点,为室内可见光信道模型提供新的研究方案。     

输出不可量测非线性系统的神经模型参考自适应控制

该文针对被控对象输出不可量测的非线性系统,引入一个便于在线辨识的扩展神经网络模型,提出一种基于前馈-反馈结构的神经网络模型参考自适应控制方法。给出了具有全局收敛性的网络训练算法,并分析了控制系统的稳定性。仿真结果表明该控制方法是有效的,而且对网络初始权值的选取及被控对象特性参数的扰动都具有良好的鲁棒性。     

A new training strategy for neural network using shuffled frog-leaping algorithm and application to channel equalization

This paper makes use of shuffled frog-leaping algorithm (SFLA) as a training algorithm to train multi-layer artificial neural network (ANN). Next, The SFLA ANNs are used for channel equalization. We, in this paper, also introduce SFLA for channel equalization that is formulated as an optimization problem. In short, this paper introduces a novel strategy for training of ANN and also proposes two novel approaches for channel equalization problem using shuffled frog-leaping algorithm (SFLA). The proposed strategies are tested both in time-invariant and time varying channels and interestingly yield better performance than contemporary approaches as evidenced by simulation results.     

地基微波辐射计分段正交神经网络方法遥感阴天大气

标准的正交神经网络方法利用一组自然正交函数将大气剖面展开,并采用神经网络的方法估计其正交函数的系数,该方法能够充分融合云底高度信息,提高阴天大气的反演精度.在此方法基础上,文章提出一种分段正交神经网络方法,即将大气剖面分为若干段,不同段采用不同的正交系数进行拟合,再利用神经网络估计其正交系数矩阵.结果表明,用此系数矩阵拟合阴天大气剖面较标准算法的相关系数更大,反演精度也更高.同时,利用青岛站微波辐射计实测数据验证了所提方法的有效性.     

Parallel data transmission for navigation satellite network with agility link

The topology of the navigation satellite network with agility link evolves with chain‐building slots periodically. The navigation data are transmitted in a “store, carry and forward” manner without requiring end‐to‐end paths. Thus, the navigation satellite network with agility link can be considered as a deterministic delay tolerant network (DTN). The data transmission of navigation satellite network is limited by chain‐building slot and bandwidth. To utilize network transmission link and communication bandwidth fully, this paper proposed an algorithm called parallel data transmission for the navigation satellite network with agility link. Based on the network characteristics, the algorithm is designed with polling pattern of link establishment and formulated as parallel data transmission problem based on deterministic scheduling DTN. Then, we design the method based on Dynamic Game from the view of data, while the Dynamic Programming method is proposed referring to Linear Programming–based method for optimization. Finally, the results were compared with three methods using the standard network routing metrics and showed that the proposed algorithm is better than the existing methods.     

Network coding based on hyper‐edge decomposition for wireless network with link failures

This paper is concerned with the construction of network coding for wireless network with link failures. Based on hyper‐edge decomposition, for the wireless network, we first construct the hyper‐edge graph, where each node represents one hyper‐edge (consisting of multiple adjacent edges transmitting the same information) of the wireless network. Then we present a heuristic coloration method for the hyper‐edge graph, and a network coding vector allocation scheme based on maximum distance separable code is proposed to effectively overcome some link failures. Copyright © 2014 John Wiley & Sons, Ltd.     

基于卷积神经网络的随机因子重采样图像检测

图像重采样检测是图像取证领域的重要任务,其目的是检测图像是否经过重采样操作。现有的基于深度学习的重采样检测方法大多只针对特定的重采样因子进行研究,而较少考虑重采样因子完全随机的情况。本文根据重采样操作中所涉及的插值技术原理设计了一组高效互补的图像预处理结构以避免图像内容的干扰,并通过可变形卷积层和高效通道注意力机制(efficient channel attention, ECA)分别提取和筛选重采样特征,从而有效提高了卷积神经网络整合提取不同重采样因子的重采样特征的能力。实验结果表明,无论对于未压缩的重采样图像还是JPEG压缩后处理的重采样图像,本文方法都可以有效检测,且预测准确率相比现有方法均有较大提升。     

On the development of adaptive hybrid active noise control system for effective mitigation of nonlinear noise

The presence of nonlinearities as well as acoustic feedback deteriorates the cancellation performance of the conventional filtered-x LMS (FxLMS) algorithm based active noise control (ANC) systems. With an objective to improve the performance, a novel filtered-su LMS (FsuLMS) algorithm based ANC system which employs a convex combination of an adaptive IIR filter with a functional link artificial neural network (FLANN) is proposed in this paper. The corresponding learning algorithm of the ANC system is derived and used in the simulation study for performance evaluation. Simulation study reveals enhanced performance of the proposed system over that of its component filters.     

大量目标识别的光学相关与神经网络融合方法

本文基于光学相关与神经网络学习算法和编码聚类原理的融合 ,提出一种大数量目标的高效率分类识别方法 ,并以 67个字符 ,其中包括2 6个英文字母、1 0个数字和 31个省和直辖市的简称汉字的分类识别作为实例 ,给出详细的原理说明和结果。     

自适应RBF神经网络在CDMA移动通信上行链路功率控制中的应用研究

提出了一种自适应RBF神经网络功率控制方案。详细研究了该网络在DS-CDMA通信中，进行上行链路闭环功率控制(基于信扰比(SIR))的应用理论，给出了该网络参数的计算方法。最后用计算机仿真法模拟出该控制器的运行性能。结果表明基于SIR的自适应RBF神经网络功率控制器能自适应地调整移动台的发射功率，使基站接收信号的信扰比始终非常接近于一个常数，且有比定步长功率控制更小的SIR跟踪误差，从而可以降低接收信号的中断概率、提高信道容量。