A data distribution scheme for VANET based on fountain code

Data dissemination is one of the applications used to provide infotainment to the end-users in vehicular ad hoc networks (VANET). During this process, the vehicles receive the data broadcast by the RoadSide Unit (RSU). However, it is difficult for vehicles to collect the complete content within the communication range of one RSU when the vehicle moves at a high speed and the amount of broadcast data is large. To solve this problem, a multi-RSU cooperative data distribution scheme based on fountain code (MRFC) is proposed in this paper. The source data are encoded by fountain code and poured into the VANET by multiple cooperative RSUs, then the vehicles in the area share data packets through the V2V resource compensation method, so that all vehicles can obtain enough encoded packets to reconstruct the source data. To improve channel resource utilization and reduce delivery delays, the RSUs use fuzzy logic to determine the number of fountain code packets according to their locations, the speed and density of surrounding vehicles. The experimental results show that on the premise of ensuring the delivery rate, the proposed scheme can reduce the delivery delay by 30–50%, and achieve a significant improvement in performance.

     

基于H表示的时变随机Markov 跳跃系统的能观性

研究时变连续和离散随机Markov 跳跃系统(SMJSs) 的能观性问题. 基于H-表示方法将时变SMJSs 转化为等价的时变线性系统, 根据线性系统理论得到时变连续和离SMJSs 的能观性Gramian 矩阵判据. 数值仿真表明了所得结论的正确性.     

融合双层信息的显著性检测

目的 针对已有工作在颜色及结构显著性描述方面的缺陷,提出一种新的图像显著性检测方法。方法 本文方法在不同的图像区域表达上从颜色与空间结构角度计算图像的显著性,充分考虑图像的特征与像素聚类方式之间的适应性。首先,根据颜色复杂度、边缘与连通性等信息,将图像从像素空间映射到双层区域表示空间。然后,根据两个层次空间的特性,与每个图像区域的边界特性,计算图像的结构和颜色显著度。最后,由于不同图像表示中的显著性信息存在互补性,将所有这些信息进行融合得到最终的显著性图。结果 在公认的MSRA-1000数据集上验证本文方法并与目前国际上流行的方法进行对比。实验结果表明,本文方法在精确率、召回率以及绝对误差(分别为75.03%、89.39%、85.61%)等方面要优于当前前沿的方法。结论 提出了一种融合双层信息的显著性检测算法。根据图像本身信息控制区域数目构建图像双层表示,提高了方法的普适性;利用图像不同层次的特性从不同角度计算显著性,增强了方法鲁棒性。     

灰狼优化的混合参数多分类孪生支持向量机

孪生支持向量机(TWSVM)是在支持向量机(SVM)的基础上产生的一种高效二分类算法,由于现实中存在的问题大多数是多分类的,将二分类孪生支持向量机扩展到多分类孪生支持向量机(MTWSVM)是非常重要的。目前常用的MTWSVM一般是基于“一对一”策略,但该策略中各子分类器都采用相同的惩罚参数以及核参数,忽略了不同子分类器之间的差异,不能使其发挥最好的作用。通过提出一种基于混合参数的多分类孪生支持向量机(MP-MTWSVM),为不同的子分类器选取合适的参数,保持分类器的多样性,进而根据“一对一”策略构建MTWSVM。TWSVM本就面临着参数难确定的问题,而MP-MTWSVM算法又引入了大量的参数,通过灰狼算法(GWO)对MP-MTWSVM的参数进行寻优,进一步提出了基于灰狼优化的混合参数多分类孪生支持向量机(GWO-MP-MTWSVM)。通过实验表明,GWO可以快速找到各子分类器的最优参数,并进一步提升了算法的准确率。     

Research on end hopping and spreading for active cyber defense

Inspired by the spread spectrum technology for communications,the concept of end spreading was proposed to represent a piece of information of the data transmission with a sequence of multiple end information,of which each piece of end information was irrelevant to the information it conveys.Thus the covert data transmission can be performed.Further,an active cyber defense model of end information hopping and spreading was presented,in which the hopping strategy was separated from the synchronization strategy.The synchronization was accomplished by means of end information spreading for synchronous authentication of both parties,which can solve the high-speed hopping synchronization problem with high concealment requirements.The mode of generation,transmission and authentication of the spreading sequence,and the data migration strategy in the end hopping and spreading model were described in detail,and the security performance and synchronization performance were analyzed and verified experimentally.Theoretical analysis and experimental results show that the cyber defense model of end information hopping and spreading has improved the availability and confidentiality of network services under high-speed hopping and has good anti-attack performance,which is of great significance for the proactive defense application of high intensity confrontation.     

涡街提频振荡水柱驱动压电发电理论分析

海洋波浪能频率低(0.3～1 Hz),海浪进入振荡水柱气室压缩空气形成高速气流。在高速气流通道内放置绕流圆柱钝体,采用卡门涡街效应提高波浪能振荡水柱采集气室气压激振频率,实现高频驱动压电发电,提高海洋能量转换效率。推导波浪进入气室形成振荡的水气动力转换理论,计算了气室初级压强和流速。分析了低频高压气流经钝体形成卡门涡街高频涡流激振提频过程,计算出气压作用于钹型压电发电结构输出的电量。研究气体流速、钝体等系统参数与输出能量的关系。计算结果表明,周期为0.65～1.1 s的海浪进入气室经提频作用于钹型压电发电结构,稳定输出电能可达70～80 mW,为新型波浪能采集技术提供了理论基础。     

高分五号全谱段光谱成像仪影像数据质量评价研究

高分五号卫星上搭载的我国自主研发的全谱段光谱成像仪是从可见光到长波红外光谱范围的星载多光谱成像仪,具有广泛的应用前景。对卫星影像进行质量评价,既是对遥感卫星是否满足设计指标的验证与检验,也可以为影像的处理与应用提供参考。利用信噪比、清晰度、信息量、辐射不均一性4个指标对高分五号全谱段光谱成像仪进行影像质量评价,并以美国陆地卫星Landsat 8影像为参考进行对比分析。结果表明:高分五号卫星全谱段光谱成像仪短波红外谱段的信噪比(320.44~388.42)略高于可见近红外谱段(208.24~238.03);近-短波红外谱段的清晰度(0.82~0.91)要高于其余谱段,尤其是长波红外谱段(0.01~0.21);可见短波红外谱段的信息量(9.01~9.97)要高于中长波红外谱段(5.71~8.31);所有12个谱段的辐射不均一性均小于2%。与Landsat 8的比较结果表明:在清晰度方面,全谱段光谱成像仪可见近红外谱段优于Landsat 8,其他谱段接近Landsat 8;在信息量方面,可见短波红外谱段与Landsat 8比较接近,但是B11、B12两个分裂窗谱段差距较大,分别相差5.23和5.61;在信噪比方面GF-5 VIMI仍有待进一步改善,又以B1、B2、B63个谱段落后Landsat 8最大,分别相差280.41、226.84和151.92。     

基于形变模型的多角度三维人脸实时重建

摘 要：采用人脸特征点调整三维形变模型的方法应用于面部三维重建,但模型形变的计 算往往会产生误差,且耗时较长。因此运用人脸二维特征点对通用三维形变模型的拟合方法进 行改进,提出了一种视频流的多角度实时三维人脸重建方法。首先利用带有三层卷积网络的 CLNF 算法识别二维特征点,并跟踪特征点位置;然后由五官特征点位置估计头部姿态,更新 模型的表情系数,其结果再作用于 PCA 形状系数,促使当前三维模型发生形变;最后采用 ISOMAP 算法提取网格纹理信息,进行纹理融合形成特定人脸模型。实验结果表明,该方法在 人脸重建过程中具有更好的实时性能,且精确度有所提高。     

基于点云数据的三维目标识别和模型分割方法

三维模型的深度特征表示是三维目标识别和三维模型语义分割的关键和前提,在 机器人、自动驾驶、虚拟现实、遥感测绘等领域有着广泛的应用前景。然而传统的卷积神经网 络需要以规则化的数据作为输入,对于点云数据需要转换为视图或体素网格来处理,过程复杂 且损失了三维模型的几何结构信息。借助已有的可以直接处理点云数据的深度网络,针对产生 的特征缺少局部拓扑信息问题进行改进,提出一种利用双对称函数和空间转换网络获得更鲁棒、 鉴别力更强的特征。实验表明,通过端到端的方式很好地解决缺少局部信息问题,在三维目标 识别、三维场景语义分割任务上取得了更好的实验效果,并且相比于 PointNet++在相同精度的 情况下训练时间减少了 20%。     

Reliability and cost efficiency analysis for deep space habitat network based on Nash bargaining solution

Focusing on the deep space habitat network requirements of high reliability and expensive links,two impact factors to the network performance were regarded as a multi-objective optimization problem,which were network reliability and cost efficiency model,and the relations between the two factors were formalized.Thereafter,the Nash model of cooperative game was used to make a fairness compromise to the network reliability and cost efficiency,and the Nash bargaining solution set was acquired through the Matlab platform with CVX.Further,through changing the threat value dynamically and performing the repeated game,the Pareto optimal solution could be obtained which can ensure the fairness of reliability and cost efficiency simultaneously.Finally,the Nash bargaining method was applied to a typical deep space habitat network topology,and described in detail how to set up the link numbers and link capacity in order to reduce the cost when the network traffic and the link capacity were given.The proposed work is important for the deployment and optimization of the deep space habitat network.