MULTI-SCALE COHERENT STRUCTURES IN TURBULENT BOUNDARY LAYER DETECTED BY LOCALLY AVERAGED VELOCITY STRUCTURE FUNCTIONS

The time sequence of longitudinal velocity component at different vertical locations in turbulent boundary layer was finely measured in a wind tunnel. The concept of coarse_grained velocity structure functions, which describes the relative motions of straining and compressing for multi_scale eddy structures in turbulent flows, was put forward based on the theory of locally multi_scale average. Based on the consistency between coarse_grained velocity structure function and Harr wavelet transformation,detecting method was presented, by which the coherent structures and their intermittency was identified by multi_scale flatness factor calculated by locally average structure function. Phase_averaged evolution course for multi_scale coherent eddy structures in wall turbulence were extracted by this conditional sampling to educe scheme. The dynamics course of multi_scale coherent eddy structures and their effects on statistics of turbulent flows were studied.

Baroclinic pressure gradient difference schemes of subtracting the local averaged density stratification in sigma coordinates models

Much has been written of the error in computing the baroclinic pressure gradient (BPG) with sigma coordinates in ocean or atmospheric numerical models. The usual way to reduce the error is to subtract area-averaged density stratification of the whole computation region. But if there is great difference between the area-averaged and the local averaged density stratification, the error will be obvious. An example is given to show that the error from this method may be larger than that from no correction sometimes. The definition of local area is put forward. Then, four improved BPG difference schemes of subtracting the local averaged density stratification are designed to reduce the error. Two of them are for diagnostic calculation (density field is fixed), and the others are for prognostic calculation (density field is not fixed). The results show that the errors from these schemes all significantly decrease.

Generalized Sampling Series Approximation of Random Signals from Local Averages

Signals are often of random character since they cannot bear any information if they are predictable for any time t, they are usually modelled as stationary random processes .On the other hand, because of the inertia of the measurement apparatus, measured sampled values obtained in practice may not be the precise value of the signal X(t) at time tk (k∈Z), but only local averages of X(t) near tk. In this paper, it is presented that a wide (or weak ) sense stationary stochastic process can be approximated by generalized sampling series with local average samples.

多视野精细分析下的弱监督目标定位算法

针对多尺度目标定位精度较差,难以捕获完整目标边界的问题,设计一种多视野精细分析模块并融入通道与空间注意力机制抑制背景噪声的干扰,获取多尺度目标的高分辨率特征。利用随机特征选取模块获取特征图随机位置的组合,聚合多个位置图获取最具辨别性的位置及其它位置的信息,融合浅层生成的类激活图与聚合类激活图获取细粒度位置信息,捕获完整的目标边界。与现有的弱监督定位方法相比,在解决多尺度目标定位效果差和局部最优问题上具有一定的优势。

基于多任务神经网络的水下震源定位方法研究

针对传统水下被动定位算法传感器阵列体积大、布设难、环境失配等问题,提出一种基于多任务神经网络的水下震源定位方法。该方法通过仿真浅海声场数据集,利用声信号到达各个测量传感器的相对时间差,结合深度学习方法,在多任务卷积神经网络MTL-CNN(Multi-task Convolutional Neural Network)和Attention-UNet结构基础上,设计MTL-Attention-UNet神经网络模型,对水下震源的距离和深度进行联合估计。仿真结果表明:用MTL-Attention-UNet模型对水下震源进行定位的平均绝对误差比MTL-CNN网络模型小,定位性能更好。

局部平均随机过程随机场采样定理

随机信号采样定理的相关研究几十年来一直是采样定理主流方向之一,在理论和实际应用两方面都受到信息领域和数学领域的广泛关注。近年来,采样理论在随机过程和随机场的局部平均采样方面得出了一系列研究成果。首先,分析了采样定理在应用中的缺陷和局限性。其次,对近几年在随机过程、高维随机过程、随机场、时空随机场局部平均采样定理方面的发展和相关结论做了系统介绍。最后,对基于混合范数的时空随机场采样逼近、采样形式、信号形式、范数形式等问题进行了展望。结果表明:局部平均的采样方式提高了采样的精度和逼近效果,随机场局部平均采样理论在海浪参数反演和海洋污染监测方面有重要应用。特别是基于混合范数的时空随机场采样理论的研究不仅具有重要的理论价值,而且对于工程领域也有应用价值。

基于四元数表示的GraphSLAM位姿初始化算法

位姿估计问题本质上是一个非凸且NP-难问题,难以求解的原因在于位姿中包含特殊正交群,且往往拥有一个较差的初始值。针对此问题,提出了一种基于图的即时定位与建图技术位姿初始化算法。首先,使用两阶段算法将位姿估计问题构建的最大似然估计方程重新拆分为两个独立的子问题,即基于四元数表示的旋转平均问题和无约束的平移向量平均的线性方程问题。然后,通过迹松弛方式将非凸问题转化为线性方程问题。最后,结合平移向量计算出初始化的位姿。仿真实验表明,在高噪声的数据集中,所提方法比基于生成树的目标函数值降低了至少1个数量级,能够鲁棒地为位姿估计问题提供一个良好的初始值。

直立宽大结构局部冰压力幅值研究

为了明确极区直立宽大结构局部冰压力幅值及压力面积曲线的影响因素,本文首先对ISO 19906规范中的局部冰压力计算方法进行了分析,发现其局限性较大。其次,基于离散元方法模拟不同海冰条件与多种尺度结构的相互作用,发现局部冰压力幅值随冰厚的增大而明显增大,随结构宽度的增大而减小;局部冰压力幅值及其标准差随宽厚比的增加而减小,表现出明显的尺寸效应。最后,采用方形区域平均法研究了冰厚和宽度两种尺寸因素对压力-面积曲线参数的影响,选取某种相似工况与ISO规范进行对比。验证了离散元模拟方法处理局部冰压力-面积曲线的合理性,可为平台的抗冰性能分析和结构设计提供一定依据和参考。

融合有效掩膜和局部增强的遮挡行人重识别

在监控系统中行人经常会被各种障碍物遮挡,使得遮挡行人重识别仍然是一个长期存在的挑战。最近一些基于Transformer和外部语义线索的方法都改善了特征的表示和相关性能,但仍存在表示弱和语义线索不可靠等问题。为解决上述问题,提出了一种基于Transformer的新方法。引入了一种有效的掩膜生成方式,可靠的掩膜可以使模型不依赖外部语义线索并实现自动对齐。提出了一种基于平均注意力分数的序列重建模块,可以更有效地关注前景信息。提出了局部增强模块,获得了更鲁棒的特征表示。比较了所提方法和现有的各种方法在Occluded-Duke,Occluded-ReID,Partial-ReID,Market-1501数据集上的性能。Rank-1准确率分别达到了72.3%、84.8%、86.5%和95.6%,mAP精度分别为62.9%、83.2%、76.4%和89.9%,实验结果表明所提模型性能较其他先进网络有所提升。

ELASTIC WAVE LOCALIZATION IN TWO-DIMENSIONAL PHONONIC CRYSTALS WITH ONE-DIMENSIONAL QUASI-PERIODICITY AND RANDOM DISORDER

The band structures of both in-plane and anti-plane elastic waves propagating in two-dimensional ordered and disordered （in one direction） phononic crystals are studied in this paper. The localization of wave propagation due to random disorder is discussed by introducing the concept of the localization factor that is calculated by the plane-wave-based transfer-matrix method. By treating the quasi-periodicity as the deviation from the periodicity in a special way, two kinds of quasi phononic crystal that has quasi-periodicity （Fibonacci sequence） in one direction and translational symmetry in the other direction are considered and the band structures are characterized by using localization factors. The results show that the localization factor is an effective parameter in characterizing the band gaps of two-dimensional perfect, randomly disordered and quasi-periodic phononic crystals. Band structures of the phononic crystals can be tuned by different random disorder or changing quasi-periodic parameters. The quasi phononic crystals exhibit more band gaps with narrower width than the ordered and randomly disordered systems.

RSS-Distance Rationalization Procedure for Localization in an Indoor Environment

The computational capabilities of off-the-shelf wireless sensors networks presents a limitation when more complex forms of localization algorithms are employed for location estimation purposes, particularly in an indoor environment. Range-free algorithms rely on Received Signal Strength (RSS) from sensors that are location aware (anchor nodes) as the major means of distance estimation. This paper presents a non-site specific algorithm for better estimating RSS relationship with distance. By employing a unique form of rationalization of raw RSS with respect to distance using the proposed algorithm, it is possible to enhance the reliability of RSS when employed in indoor Localization Algorithms. Consequently, this paper presents an innovative RSS-Distance rationalization algorithm for localization of objects in an indoor environment. The paper compared the proposed algorithm with Simple Moving Average (SMA) algorithm due to the wide applicability and ease of manipulation of SMA. The analysis of the proposed algorithm and SMA shows that the proposed algorithm better modifies RSS for more accurate position estimation in an indoor environment.

The Non—local Thermodynamical Equilibrium Effects on Opacity

Based on the detailed configuration accounting (DCA) model, a method is developed to include the resonant photoionization and the excitation-autoionization in the non-local thermodynamical equilibrium (NLTE) average atom(AA) model. Using this new model, the mean charge states and the opacity are calculated for NLTE high-Z plasmas and compared with other results. The agreement w ith AA model is poor at low electron density. The present results agree well with those of DCA model within 10%. The calculations show that the NLTE effects on opacity are strong.

LOCAL HOLDUPS AND LIQUID INTERSTITIAL VELOCITY IN LARGE-DIAMETER THREEPHASE FLUIDIZED BED

Radial profiles of local gas and solid holdups and liquid interstitial velocity were measured in a largediameter three-phase fluidized bed, and a systematic study was carried out in comparing the present data with those obtained in small-diameter beds. Radial profiles of local gas holdup were found to be parabolic; the distribution of local solid particle holdup could be expressed by Eq. (9); liquid interstitial velocity was well described by the modified circulating flow model.

长江流域极端降水与当地平均气温的响应关系分析

研究极端降水的非平稳性及对气温的响应有助于应对相关灾害事件带来的消极影响。基于长江流域1961-2020年降水和气温格点数据,建立广义极值模型以确定当日(T_(0))和前1~9 d(T_(1-9))当地平均气温是否会引起极端降水的非平稳性,然后通过等样本箱元法、分段线性回归法和指数回归法探讨极端降水对T_(0)和T_(1-9)的响应。结果表明:长江流域99.7%格点的极端降水是非平稳的,T_(0)和T_(1-9)共驱动74.4%格点的极端降水向非平稳演变,且T_(1-9)对极端降水的影响更大。流域极端降水随T_(0)升高主要呈低温时增加而高温时减少的峰值型结构,且峰值点气温以青藏高原东缘为界西低而东高并分别集中于9℃和24℃左右;同时,极端降水随T_(0)的变化率介于(-12.3%~53.6%)/℃之间,并以四川盆地为中心向四周呈超CC(Clausius-Clapeyron)、类CC和次CC变率的分布格局。极端降水随T_(1-9)上升在流域西部主要呈增加趋势而在中东部主要为峰值型结构,且峰值点在降水强度最大时出现在25℃附近;此外,极端降水随T_(1-9)的变化率由西至东从超CC变率过渡到类CC和次CC变率,并处于(-3.7%~33.8%)/℃之间且集中于4%/℃和10%/℃附近。随着降水极端性的增强,极端降水对气温变化的敏感性降低从而使其变化率的范围缩小;此外,超CC变率可能与对流降水和潜热释放等相联系,而峰值型结构或许与降水的冷却作用以及反气旋活动等有关。

海量文献数据无线传感DV-Hop定位算法设计

现有的数据无线传感定位算法无法获取锚节点与待定位节点之间的最小跳数,导致定位偏差较大,且可应用的网络环境单一。针对海量文献数据,设计新的无线传感DV-Hop定位算法。利用文献数据位置信息表获取锚节点至待定位文献数据未知节点的最小跳数,利用所确定的最小跳数确定无线传感器网络中每跳的平均距离,选取极大似然估计法利用每跳的平均距离确定待定位未知节点的定位结果。通过修正锚节点间跳数、修正未知节点跳数以及修正未知节点平均跳距三部分优化DV-Hop定位算法,提升海量文献数据定位精度。实验结果表明,修正系数为2.05时,文献数据定位精度较高;定位精度伴随锚节点覆盖率的提升有所降低;不同网络场景时,所研究算法的定位精度均高于0.94。

城投债为何持续增长——基于地方政府网络博弈视角的分析

本文基于地级市层面的城投债发行数据和地理网络数据,利用网络博弈分析方法,考察了城投债持续增长的网络效应。研究结果发现,城投债发行规模显著地受到地理网络的影响。城投债增长的地理网络效应以局部平均效应为主,即相邻城市的城投债均值增长会带动本地城投债增长。在同一个省份内部,城投债发行存在互补效应,本地城投债规模随着省内相邻城市城投债总和增加而增长。给定地方政府的城投债发行与相邻城市债务相关且存在局部平均效应,为了控制地方政府隐性债务快速增长,相比于约束个别地方政府行为,中央政府更应该实施群体性的规范来加强地方政府债务管理,避免地方政府之间相互攀比,进而降低整体债务快速增长的风险。

基于一种新的联邦优化算法的信用风险预测方法

ResNet深度神经网络模型与联邦学习融合建模,实现信用风险预测并保证用户数据的隐私安全。模型准确率和精准度可达89.67%和94.22%。首次采用联邦优化算法-局部全局联合平均算法(LG-FEDAVG)优化训练过程,对贷款用户的信用风险建模。研究表明,LG-FEDAVG算法在保证信用风险预测效果的情况下产生的通信成本占通信总参数的3.22%。

Approximation of weak sense stationary stochastic processes from local averages

We show that a weak sense stationary stochastic process can be approximated by local averages. Explicit error bounds are given. Our result improves an early one from Splettstosser.

一种基于被动声呐宽带空间谱的自动检测算法

被动声呐宽带目标检测中,基于固定门限的检测不能适应波动背景、对强弱目标共存场景检测时易检出旁瓣或漏检弱目标,此外检测结果中目标占据多个波束易使精确测向产生引导波束选取偏差。针对上述问题,基于宽带空间谱,将排序截断平均算法、局部峰值检测和波束内插算法结合,通过理论和蒙特卡洛实验分析对门限估计值有影响的参数,设计参数自动选取方案,形成检测门限自适应方法及自动检测与精确测向一体化方案。仿真与海试实验结果验证了所提算法的有效性,有较强工程应用价值。

Uniform Truncation Error for Shannon Sampling Expansion from Local Averages

Let B^pΩ,1≤Р≤∞,be the set of all bounded functions in L^p（R）which can be extended to entire unctions of exponential typeΩ. The unitbrm bounds for truncation error of Shannon sampling expansion fromlocal averages are obtained for functions f∈BpΩwith the decay condition f（t）≤A/t^δ,t≠0,where A and δare positive constants. Furthermore we also establish similar results for non-bandlimit functions in Besov classes with the same decay condition as above.