Evaluating Efficiency of Decision Making Units with Network Structures

This paper considers the problem of evaluating efficiency of Decision Making Units （DMUs） with network structures of divisions by the Data Envelopment Analysis （DEA） model. All divisions in the network are under a decentralized authority organiza- tion. That is, each division in a decision making unit has its own authority to adjust its input and output. By incorporating the division operations in the DEA model, we discuss the sufficient and necessary conditions for a DMU to be network efficient in series structure and general structure respectively.     

一种变阶数自适应滤波算法

通过对回声抵消的实际应用环境进行分析,提出一种新的动态改变自适应滤波器阶数的算法。该算法在权系数的长度与失调误差信号之间建立一种非线性函数关系,具有在环境突变阶段,保持较高的滤波器阶数,而在稳态阶段自动减少阶数的特点,可以极大地降低自适应滤波器算法的运算量。理论分析和计算机仿真结果表明,该算法具有良好的收敛性能和跟踪性能,稳态误差的水平与传统算法保持一致。     

带落角约束的再入机动弹头的变结构导引律

针对再入机动弹头垂直打击目标的要求,研究了具有末端落角约束的变结构导引律.该导引律包括俯冲平面内的制导方程和转弯平面内的制导方程,通过采用准滑动模态控制削弱抖振的影响.通过数学仿真,分析了四个制导参数对制导效果的影响,探讨了利用遗传算法离线优化制导参数的可行性,检验了最优制导参数的制导效果对不同初始状态偏差的敏感性.仿真结果表明,与最优导引律相比,变结构导引律具有较强的鲁棒性,在干扰影响下仍能保持较高的制导精度.     

DDFS spurious signals due to amplitude quantization in absence of phase-accumulator truncation

Spurious signals in direct digital frequency synthesizers (DDFSs) are partly caused by amplitude quantization and phase truncation, which affect their application to many wireless telecommunication systems. Thee signals are deterministic and periodic in the time domain, so they appear as line spectra in the frequency domain. Two types of spurious signals due to amplitude quantization are exactly formulated and compared in thetime and frequency domains respectively. Then the frequency spectra and power levels of the spurious signals due to amplitude quantization in the absence of phase-accumulator truncation are emphatically analyzed, and the effects of the DDFS parameter variations on the spurious signals are thoroughly studied by computer simulation. And several important conclusions are derived which can provide theoretical support for parameter choice and spurious performance evaluation in the application of DDFSs.     

Uniform optimal-order estimates for finite element methods for advection-diffusion equations

This article summarizes our recent work on uniform error estimates for various finite element methods for time-dependent advection-diffusion equations.     

商业智能在社保系统中应用的实现框架

作为联机分析处理(OLAP)和数据挖掘、数据仓库等技术的综合应用,商业智能(BI)是目前在国内外企业界和软件开发界受到广泛关注的一个研究方向.论文在分析BI的概念、技术支柱和三维框架基础上,结合社会保障数据的特点.阐述了社保系统中实施BI应用的必要性,最后,给出了社保系统中BI应用的框架和技术方案.     

基于支持向量机的体数据分类算法研究

在由工业CT图像构成的体数据中,低密度材质的灰度与背景及伪影的灰度接近,影响了分类的准确性,进而影响体绘制的结果.针对这一问题,提出了一种基于支持向量机的体数据分类算法.首先以体素的灰度、梯度、局部直方图的熵和矩为样本特征进行训练,得到决策函数,并通过决策函数对体数据进行类别标记,再将具有相同标记的体素的灰度值变换到指定的灰度区间,最后利用阻光度传递函数对体素进行阻光度赋值,完成体数据的分类.实验结果表明,所提出的算法较好的实现了体数据的分类,体绘制结果清晰,且能够实现试件的模拟拆卸.     

基于发射分集的MIMO雷达相参信号处理方法

针对多个运动的慢起伏目标情况,研究了发射分集多入多出(multiple input multiple output,MIMO)雷达的相参信号处理方法.提出了一种重复利用回波数据的目标参数估计算法.发射分集MIMO雷达的发射阵元间距满足空间分集条件,其接收端采用阵元间距为半波长的均匀线阵.按距离门对其回波信号进行脉压-Capon谱估计,可提取出各目标对应不同发射阵元的初步位置信息.然后利用回波数据及目标的初步位置信息,把同一目标对应不同发射阵元的脉压数据矢量经多普勒频移和相位补偿后相加,再进行Capon谱估计,获得高精度的目标方位角估计值.仿真结果表明,该算法可对雷达观测区域内的多个动目标进行高精度定位.     

Analysis of electrical performances of planar active phased array antennas with distorted array plane

The influence of the distorted plane of the active phased array antenna on the electromagnetic performance is of great significance to the research on and development of the high-performance antennas. On the bent and bowl-shape distortion, the model is established of the relationship between the electromagnetic performance and the position error of the radiated elements. The method is presented of analyzing the far-field pattern of the distorted rectangular active phased array antenna. The analysis results of a planar phased array antenna with different distortions grades prove the validity of the model. Therefore, by the method, the antenna designers may set the reasonable requirement on the structural tolerance in manufacturing antenna.     

基于堆栈式LSTM网络的通信辐射源个体识别

针对现有通信辐射源个体识别方法预处理过程复杂及特征提取较难的问题,提出了一种基于堆栈式长短期记忆(long short-term memory, LSTM)网络的辐射源个体识别算法。该算法直接使用IQ时间序列信号训练LSTM网络,即可实现对通信辐射源个体的高效识别,避免了复杂的信号预处理过程。为使LSTM网络能更好地适用于通信辐射源个体识别,利用3层LSTM网络提取辐射源深层特征,并通过实验优化了网络参数。然后对该算法的实际应用泛化性进行了实验探究,结果表明该算法在其他辐射源数据集上也取得了较好的效果。最后,通过实验对算法进行了验证,结果表明相比于传统算法,在样本数较多时,该算法的识别准确率可以达到98%,而且简单快速智能,便于工程化与实用化。