基于GAN的通信干扰波形生成技术

现有通信干扰方法, 通常基于通信侦察中获取的目标信号特征进行干扰决策, 选取合适的干扰波形实施干扰, 难以应对目标信号特征未知或参数动态变化的情况。为此, 提出一种基于生成对抗网络(generative adversarial networks, GAN)的通信干扰波形生成技术, 运用GAN直接提取目标信号的潜在特征, 并生成与目标信号特征相似的干扰波形。在介绍GAN原理的基础上, 首先设计网络模型, 并对学习率进行优化, 使GAN更适用于时间序列通信干扰波形的生成。然后通过对不同类型和参数的通信信号进行干扰波形生成实验, 验证了该技术的泛化性。最后进行干扰效果对比试验, 结果表明, GAN生成的干扰波形干扰效果能够逼近最佳干扰效果。     

基于规则分组的DFA正则表达式匹配算法

模式匹配因误报率低和漏报率低被入侵检测所采用.在使用正则表达式构造DFA时,因状态爆炸导致匹配算法需要较多的存储空间和运行时间,算法效率低下,采用规则分组后,可以在一定程度上抑制状态爆炸问题.根据缓存中的历史记录对正则表达式进行分组,既能利用规则分组减少状态总数,抑制状态爆炸,又能减少因每次重新构建DFA所带来的开销,提高了匹配效率,有利于提高入侵检测的实时性、准确性和高效性.     

科技创新、科技产业与科技金融时空耦合分析

选取我国30个省市2008～2017年的面板数据,利用科技创新、科技产业和科技金融三系统耦合协调度模型进行测算,针对各省市三系统耦合协调度进行空间相关性检验.结果显示,我国科技创新、科技产业与科技金融整体处于中度耦合协调水平,且时间上呈现出周期波动的特征,空间上呈现出"东高西低"的态势;长三角地区始终位于H-H区域,而西北地区位于L-L区域.得出如下建议,科技金融应形成支持科技发展的长效机制;地方政府应推出因地制宜的科技金融政策;中央政府应对低度耦合发展集聚的区域给予相应的优惠补贴和鼓励措施.     

Improving the photocatalytic activity of graphitic carbon nitride by thermal treatment in a high-pressure hydrogen atmosphere

Extending visible light absorption range and suppressing the recombination of photogenerated charge carriers are always important topics in developing efficient solar-driven photocatalysts. In this study, the thermal treatment process at 400 °C in a high-pressure hydrogen atmosphere was applied to modify graphitic carbon nitride. Compared to the normal atmospheric hydrogen treatment process, this process has the merit of producing nitrogen deficient graphitic carbon nitride in high-yield. The optimal photocatalytic activity of modified graphitic carbon nitride was demonstrated by controlling the treatment duration in the hydrogen atmosphere.The changes in the crystal structure, microstructure and optical properties of carbon nitrides were investigated by several characterizations. The relationship between the photocatalytic activity and structures of graphitic carbon nitride was preliminarily established. The results obtained in this study could provide some new ways of improving the activity of graphitic carbon nitride based photocatalyst.     

Bidirectional Feedback Dynamic Particle Filter with Big Data for the Particle Degeneracy Problem

Particle Filter(PF) is a data assimilation method to solve recursive state estimation problem which does not depend on the assumption of Gaussian noise, and is able to be applied for various systems even with non-linear and non-Gaussian noise. However, while applying PF in dynamic systems, PF undergoes particle degeneracy,sample impoverishment, and problems of high computational complexity. Rapidly developing sensing technologies are providing highly convenient availability of real-time big traffic data from the system under study like never before. Moreover, some sensors can even receive control commands to adjust their monitoring parameters. To address these problems, a bidirectional dynamic data-driven improvement framework for PF(B3 DPF) is proposed.The B3 DPF enhances feedback between the simulation model and the big traffic data collected by the sensors,which means the execution strategies(sensor data management, parameters used in the weight computation,resampling) of B3 DPF can be optimized based on the simulation results and the types and dimensions of traffic data injected into B3 DPF can be adjusted dynamically. The first experiment indicates that the B3 DPF overcomes particle degeneracy and sample impoverishment problems and accurately estimates the state at a faster speed than the normal PF. More importantly, the new method has higher accuracy for multidimensional random systems.In the rest of experiments, the proposed framework is applied to estimate the traffic state on a real road network and obtains satisfactory results. More experiments can be designed to validate the universal properties of B3 DPF.     

Mechanical and corrosion properties of low-carbon steel prepared by friction stir processing

Low-carbon steel plates were successfully subjected to normal friction stir processing(NFSP) in air and submerged friction stir processing(SFSP) under water, and the microstructure, mechanical properties, and corrosion behavior of the NFSP and SFSP samples were investigated. Phase transformation and dynamic recrystallization resulted in fine-grained ferrite and martensite in the processed zone. The SFSP samples had smaller ferrites(5.1 μm), finer martensite laths(557 nm), and more uniform distribution of martensite compared to the NFSP samples. Compared to the base material(BM), the microhardness of the NFSP and SFSP samples increased by 19.8% and 27.1%, respectively because of the combined strengthening effects of grain refinement, phase transformation, and dislocation. The ultimate tensile strengths(UTSs) of the NFSP and SFSP samples increased by 27.1% and 38.7%, respectively. Grain refinement and martensite transformation also improved the electrochemical corrosion properties of the low-carbon steel. Overall, the SFSP samples had better mechanical properties and electrochemical corrosion resistance than the NFSP samples.     

基于R语言的迪基-福勒检验的蒙特卡罗模拟

迪基-福勒(Dickey-Fuller)检验是时间序列的平稳性检验中常用的一种方法;由于检验统计量的极限分布是由标准维纳过程关于轨道的积分来表达的,很难得到其密度函数的显式表达式,因而确定检验临界值非常困难,蒙特卡罗方法是解决这类问题的金钥匙;基于R语言对t统计量的平稳性检验的临界值的随机模拟程序进行了研究,填补了文献空白,其计算程序和方法对于金融工程或经济计量统计分析与研究具有广泛的指导意义。     

QT工艺对一种B-Nb低碳贝氏体钢 组织及强韧性的影响

通过室温拉伸、摆锤冲击、光学及扫描电镜研究了QT(淬火+回火)工艺对实验钢晶界比例、晶粒尺寸、碳化物析出情况及强韧性的影响，并对其变体的分布及组合方式进行了分析.结果表明:随着回火温度的升高，实验钢强度逐渐降低，塑性逐渐提高;450T和500T钢断口为准解理型断裂，贝氏体板条间析出的碳化物及较低比例的大角度晶界使得冲击韧性较差;而600T和650T钢断口为韧窝断裂，组织中大角度晶界的比例增加，有效地阻碍了裂纹的扩展.变体分析表明，450T钢变体组合方式介于Bain group和CP(close packed)group之间，而600T钢变体之间呈现较明显的CP组合方式，同一CP group内的变体取向差较大，偏折了裂纹传播路径，提高了低温韧性.     

Attack and Defense Strategies in Complex Networks Based on Game Theory

To investigate the attack and defense strategies in complex networks, the authors propose a two-player zero-sum static game model with complete information which considers attack and defense strategies simultaneously. The authors assume that both the attacker and defender have two typical strategies: Targeted strategy and random strategy. The authors explore the Nash equilibriums of the attacker-defender game and demonstrate that when the attacker's attack resources are not so significantly abundant as the defender's resources, there exists a pure-strategy Nash equilibrium in both model networks and real-world networks, in which the defender protects the hub targets with large degrees preferentially, while the attacker prefers selecting the targets randomly. When the attack resources are much higher than defense resources, both the attacker and the defender adopt the targeted strategy in equilibriums. This paper provides a new theoretical framework for the study of attack and defense strategies in complex networks.