基于可见光谱和改进注意力的农作物病害识别

基于可见光谱的农作物病害自动化识别和诊断是一个具有挑战性的研究领域,但现有基于卷积神经网络进行病害识别的研究往往利用深层网络牺牲模型参数量来提高对单一农作物病害识别的准确率,从而造成硬件资源的浪费.为提高农作物病害识别的准确率且避免深层网络的使用,该研究将注意力机制引入农作物病害识别领域,提出了一种基于可见光谱和改进注...     

UWB LOS/NLOS identification in multiple indoor environments using deep learning methods

Indoor location-aware service is booming in daily life and business activities, making the demand for precise indoor positioning systems thrive. The identification between line-of-sight (LOS) and non-line-of-sight (NLOS) is critical for wireless indoor time-of-arrival-based localization methods. Ultra-Wide-Band (UWB) is considered low cost among the many wireless positioning systems. It can resolve multi-path and have high penetration ability. This contribution addresses UWB NLOS/LOS identification problem in multiple environments. We propose a LOS/NLOS identification method using Convolutional Neural Network parallel with Gate Recurrent Unit, named Indoor NLOS/LOS identification Neural Network. The Convolutional Neural Network extracts spatial features of UWB channel impulse response data. While the Gate Recurrent Unit is an effective approach for designing deep recurrent neural networks which can extract temporal features. By integrating squeeze-and-extraction blocks into these architectures we can assign weights on channel-wise features. We simulated UWB channel impulse response signals in residential, office, and industrial scenarios based on the IEEE 802.15.4a channel model report. The presented network was tested in simulation scenarios and an open-source real-time measured dataset. Our method can solve NLOS identification problems for multiple indoor environments. Thus more versatile compare with networks only working in one scenario. Popular machine learning methods and deep learning methods are compared against our method. The test results show that the proposed network outperforms benchmark methods in simulation datasets and real-time measured datasets.     

Network-topology-adaptive quantum conference protocols

As an important application of the quantum network communication,quantum multiparty conference has made multiparty secret communication possible.Previous quantum multiparty conference schemes based on quantum data encryption are insensitive to network topology.However,the topology of the quantum network significantly affects the communication efficiency,e.g.,parallel transmission in a channel with limited bandwidth.We have proposed two distinctive protocols,which work in two basic network topologies with efficiency higher than the existing ones.We first present a protocol which works in the reticulate network using Greeberger-Horne-Zeilinger states and entanglement swapping.Another protocol,based on quantum multicasting with quantum data compression,which can improve the efficiency of the network,works in the star-like network.The security of our protocols is guaranteed by quantum key distribution and one-time-pad encryption.In general,the two protocols can be applied to any quantum network where the topology can be equivalently transformed to one of the two structures we propose in our protocols.     

A multi-timescale resource allocation algorithm based on self-learning for distributed fog radio access networks

Faced with limited network resources, diverse service requirements and complex network structures, how to efficiently allocate resources and improve network performances is an important issue that needs to be addressed in 5G or future 6G networks. In this paper, we propose a multi-timescale collaboration resource allocation algorithm for distributed fog radio access networks (F-RANs) based on self-learning. This algorithm uses a distributed computing architecture for parallel optimization and each optimization model includes large time-scale resource allocation and small time-scale resource scheduling. First, we establish a large time-scale resource allocation model based on long-term average information such as historical bandwidth requirements for each network slice in F-RAN by long short-term memory network (LSTM) to obtain its next period required bandwidth. Then, based on the allocated bandwidth, we establish a resource scheduling model based on short-term instantaneous information such as channel gain by reinforcement learning (RL) which can interact with the environment to realize adaptive resource scheduling. And the cumulative effects of small time-scale resource scheduling will trigger another round large time-scale resource reallocation. Thus, they constitute a self-learning resource allocation closed loop optimization. Simulation results show that compared with other algorithms, the proposed algorithm can significantly improve resource utilization.     

Secure collaborative cognitive radio based on chaotic modulation and compressive sensing

Cognitive radio (CR) is a wireless technology that is used to overcome the spectrum scarcity problem. CR includes several stages, spectrum sensing is the first stage in the CR cycle. Traditional spectrum sensing (SS) techniques have many challenges in the wideband spectrum. CR security is an important problem, since when an attacker from outside the network access the sensing information this produces an increase in sensing time and reduces the opportunities for exploiting vacant band. Compressive sensing (CS) is proposed to capture all the wideband spectrum at the same time to solve the challenges and improve the performance in the traditional techniques and then one of the traditional SS techniques are applied to the reconstructed signal for detection purpose. The sensing matrix is the core of CS must be designed in a way that produces a low reconstruction error with high compression. There are many types of sensing matrices, the chaotic matrix is the best type in terms of security, memory storage, and system performance. Few works in the literature use the chaotic matrix in CS based CR and these works have many challenges: they used sample distance in the chaotic map to generate a chaotic sequence which consumes high resources, they did not take into consideration the security in reporting channel, and they did not measure their works using real primary user (PU) signal of a practical application under fading channel and low SNR values. In this paper, we propose a chaotic CS based collaborative scenario to solve all challenges that have been presented. We proposed a chaotic matrix based on the Henon map and use the differential chaotic shift keying (DCSK) modulation to transmit the measurement vector through the reporting channel to increase the security and improve the performance under fading channel. The simulation results are tested based on a recorded real-TV signal as PU and Compressive Sampling Matching Pursuit (CoSaMP) recovery algorithm under AWGN and TDL-C fading channels in collaborative and non-collaborative scenarios. The performance of the proposed system has been measured using recovery error, mean square error (MSE), derived probability of detection (Pdrec), and sensitivity to initial values. To measure the improvement introduced by the proposed system, it is evaluated in comparison with selected chaotic and random matrices. The results show that the proposed system provides low recovery error, MSE, with high Pdrec, security, and compression under SNR equal to −30 dB in AWGN and TDL-C fading channels as compared to other matrices in the literature.     

Optimal design of virtual topology reconfiguration in WDM optical networks

Virtual topology of WDM optical networks is often designed for some specific traffic matrix to get the best network performance. When traffic demand imposed on WDM optical networks changes, the network performance may degrade and even become unacceptable. So virtual topology need to be reconfigured. In previous works, virtual topology is reconfigured to achieve the best network performance, in which a large number of lightpaths need to be set up or torn down. In this paper, we try to get a tradeoff between the network performance and traffic disruption (or implementing cost). The problem of virtual topology reconfiguration for changing traffic patterns is formulated as an optimization problem and a mixed integer linear programming (MILP) algorithm is presented. Numerical results show that a large cost reduction of reconfiguration can be achieved at the expense of network performance.     

A Fault-Tolerant Transmission Scheme in SDN-Based Industrial IoT (IIoT) over Fiber-Wireless Networks

Driven by the emerging mission-critical and data-intensive applications in industrial intelligent manufacturing, the software-defined network (SDN) based fiber-wireless access network (FiWi) is attracting considerable attention thanks to its capability of central control and large bandwidth. However, the heterogeneity of the network leads to new challenges, since the packet loss can be caused either by the poor channel quality of wireless links or network component failures. A novel and adaptive mechanism combining sparse random linear network coding with parallel transmission (SNC-PT) is proposed to achieve the fault-tolerance against high packet loss rate and any network element malfunction. We illustrate the benefits of using the SNC-PT mechanism to improve fault tolerance by characterizing the network performance with respect to the completion time and goodput along with its relationship to channel quality and node failures. We show that significant performance gains can be obtained in comparison with conventional uncoded transmission based on transmission control protocol (TCP). The simulation results show that the SNC-PT mechanism is fault-tolerant, while it can significantly shorten the data transmission completion time to at least 12% of the baseline and increase the goodput by about 10% compared to other coding schemes such as random linear network coding.     

Channel block-optimized spiking activity of Hodgkin-Huxley neurons on random networks

In this paper, we numerically study the effect of channel block on the spiking temporal coherence and spatial synchronization on Hodgkin-Huxley (HH) neuron networks. It is found that under sodium CB the spike coherence is badly reduced, and the synchronization can, depending on the network randomness (the fraction of random shortcuts), be either enhanced or reduced, while, under potassium CB, the spike coherence can be enhanced but the synchronization is reduced. Interestingly, for certain networks of relatively large randomness, the neuron firings can achieve the best temporal coherence at an optimal potassium CB. These results show that under certain conditions channel blocking can increase and optimize the spike coherence and the synchronization on the complex HH neuron networks, whereby the neurons would exhibit a better and the best sub-threshold signal encoding.     

基于水样类型识别的光谱COD测量方法

基于紫外吸收光谱的COD测量方法,尽管具有快速、实时、免试剂、无污染等优势。但该方法对于组分多变的水样适应性不强,构建的单一计算模型不能适用于所有待测水样类型,导致其在复杂环境下测量准确度较低,从而限制了其应用领域。本研究提出一种基于水样类型识别的测量方法。其过程包括:动态识别水样类型→自动选择相应的"吸光度(Auv)-COD"算法模型→计算COD。该方法有效提高了紫外光谱法COD测量的准确度和适用性。该研究在传统的光谱识别技术的基础上,针对COD实际测量的特点加以改进。选取水样吸光度曲线的形貌特征作为水样类型的表征参数,利用LM-BP神经网络作为识别算法。并引入了"历史数据队列"、"历史识别因子"的概念,在此基础上形成了级联的神经网络结构。该算法实现了COD测量应用中的高准确度的光谱识别,进而提高了复杂环境下COD测量的精度。大量实验测试和结果表明,与传统的光谱识别技术相比,该方法在COD测量应用中具有更好的鲁棒性和准确性。水样类型识别准确率达98%以上。同时算法结构简单,计算量小,适用于资源受限的小型化COD测量仪。当仪器在复杂多变的水环境中进行测量时,采用该算法测量得到的COD精度有显著的提高。该方法的提出为光谱COD测量法在水体组分多变场合的应用及提高其测量精度提供了技术保证,可望解决传统紫外光谱COD测量法难以适应变化和复杂水环境应用的问题。     

Perfect controlled joint remote state preparation independent of entanglement degree of the quantum channel

We construct a quantum circuit to produce a task-oriented partially entangled state and use it as the quantum channel for controlled joint remote state preparation. Unlike most previous works, where the parameters of the quantum channel are given to the receiver who can accomplish the task only probabilistically by consuming auxiliary resource, operation and measurement, here we give them to the supervisor. Thanks to the knowledge of the task-oriented quantum channel parameters, the supervisor can carry out proper complete projective measurement, which, combined with the feed-forward technique adapted by the preparers, not only much economizes (simplifies) the receiver's resource (operation) but also yields unit total success probability. Notably, such apparent perfection does not depend on the entanglement degree of the shared quantum channel. Our protocol is within the reach of current quantum technologies.     

Task offloading for vehicular edge computing with imperfect CSI: A deep reinforcement approach

This article examines a multi-user mobile edge computing (MEC) system for the Internet of Vehicle (IoV), where one edge point (EP) nearby the vehicles can help assist in processing the compute-intensive tasks. For the MEC networks, the majority of existing works concentrate on the minimization of system cost of task offloading under the perfect channel estimation, which however fails to consider the practical limitation of imperfect channel estimation (CSI) because of vehicles’ high-mobility. Therefore, the goal of our study is to reduce the delay as well as energy consumption (EC) of computation and communication with imperfect CSI, which are the two significant performance metrics of MEC network. With this aim, we first express the system cost as a form of the linear combination of the delay and EC, and then formulate the optimization problem for the system cost. Moreover, a novel deep approach is proposed, which is integrated by deep reinforcement learning (DRL) with the Lagrange multiplier to jointly minimize the system cost. In particular, the DRL algorithm is employed to obtain the capable offloading strategy, while the Lagrange multiplier is used to obtain the bandwidth allocation. The simulated results are finally presented to show that the devised approach outperforms the traditional ones.     

Black hole entropy quantization

Ever since the pioneering works of Bekenstein and Hawking, black hole entropy has been known to have a quantum origin. Furthermore, it has long been argued by Bekenstein that entropy should be quantized in discrete (equidistant) steps given its identification with horizon area in (semi-)classical general relativity and the properties of area as an adiabatic invariant. This lead to the suggestion that the black hole area should also be quantized in equidistant steps to account for the discrete black hole entropy. Here we shall show that loop quantum gravity, in which area is not quantized in equidistant steps, can nevertheless be consistent with Bekenstein's equidistant entropy proposal in a subtle way. For that we perform a detailed analysis of the number of microstates compatible with a given area and show consistency with the Bekenstein framework when an oscillatory behavior in the entropy-area relation is properly interpreted.     

基于休眠机理的三维小基站蜂窝网络能效优化

小基站通常部署在写字楼、商贸区等城市密集区域以弥补传统宏基站在覆盖和传输方面的不足.小基站的分布一般是根据高峰时的网络负荷设计的,这必然导致网络负荷较低时的资源浪费.讨论了在平均接入率和信道容量双重约束下基于休眠机理的三维小基站蜂窝网络的能效优化问题.借助泊松点过程理论推导了三维小基站网络下行信道容量和平均接入率的数学表达式.通过分析下行信道容量和平均接入率的单调性得出同时满足传输信道容量和接入率要求的最佳休眠概率.分析了小基站最大用户连接数的最佳值,通过对该参数的合理配置,可以在满足通信指标的前提下最大程度地降低网络能耗.仿真结果表明,设计的基站休眠机理可以使小基站网络的能耗下降约21%.     

Unsupervised LoS/NLoS identification in mmWave communication using two-stage machine learning framework

Identification of line-of-sight (LoS)/ non-LoS (NLoS) condition in millimeter wave (mmWave) communication is important for localization and unobstructed transmission between a base station (BS) and a user. A sudden obstruction in a link between a BS and a user can result in poorly received signal strength or termination of communication. Channel features obtained by the estimation of channel state information (CSI) of a user at the BS can be used for identifying LoS/NLoS condition. With the assumption of labeled CSI, existing machine learning (ML) methods have achieved satisfactory performance for LoS/NLoS identification. However, in a real communication environment, labeled CSI is not available. In this paper, we propose a two-stage unsupervised ML based LoS/NLoS identification framework to address the lack of labeled data. We conduct experiments for the outdoor scenario by generating data from the NYUSIM simulator. We compare the performance of our method with the supervised deep neural network (SDNN) in terms of accuracy and receiver characteristic curves. The proposed framework can achieve an accuracy of 87.4% and it outperforms SDNN. Further, we compare the performance of our method with other state-of-the-art LoS/NLoS identification schemes in terms of accuracy, recall, precision, and F1-score.     

太赫兹光谱技术在毒品检测中的应用研究

综述了近年来将太赫兹光谱技术应用于毒品检测与识别方面的研究成果：利用自主研发的可移动式小型太赫兹时域光谱仪作为实验平台,建立了含有38种纯度在90%以上的毒品太赫兹光谱数据库;用密度泛函理论进行了光谱解析;讨论了干涉以及包装物对光谱的影响;结合人工神经网络、支持向量机等方法对毒品光谱进行定性识别;同时,研究确定毒品纯度和有效成份含量的理论和实验方法。     

Story visualization of literary works

With the rapid advance in information technology, the applicability of computers has moved from the scientific field towards simulating human intelligence. We are already familiar with using computers to produce music and art and for language translation. A further use is in understanding traditional man-made products; best exemplified by literary works. In this study, we focus on enabling a computer to visualize the meaning of stories. Four world-famous plays by William Shakespeare have been chosen to demonstrate how the visualization scheme works in grasping the meaning of the stories. The scheme employs primitive keyword detection and ellipsoidal differential equations to create a visual imagery of the story. This methodology ensures uniqueness in the visualization of an individual work. In addition, color palettes obtained from pictures relevant to each story are used to enrich the consistency between the visual sense and the meaning of the story.     

Impact of direct links on Intelligent Reflect Surface-aided MEC networks

This paper studies an intelligent reflect surface (IRS) aided mobile edge computing (MEC) network, where the direct link exists in the network can assist the task transmission for computing with the help of multiple elements in the IRS. We perform the performance evaluation by instigating the impact of direct link on the outage probability. Specifically, Firstly, we analyze the system outage probability (SOP) with a different number of reflecting elements and energy consumption constraints. Moreover, we propose two selection methods for the case of multiple reflecting elements. In particular, Method I maximizes the first-hop reflecting channel while Method II maximizes the dual-hop product channel. In further, for the two different methods, we estimate the outage probability of the system by considering the reflecting channel information and providing the analytic expression of the outage probability, respectively. Finally, the numerical results verify the correctness of our results. The results show that increasing the number of reflecting elements can effectively reduce the SOP.     

基于深度信念网络的多品种玉米单倍体定性鉴别方法研究

单倍体育种技术是玉米育种新方法,该方法可有效缩短产生纯合系的周期,提高育种效率。该技术需首先挑选足量单倍体籽粒,而玉米在未加人工干预时,单倍体在混合籽粒中仅占0.05%～0.1%,即使采用生物诱导技术,单倍体籽粒数一般也不到籽粒总数的10%。高速、精准地从大量混合籽粒中挑选得到占比少于10%的单倍体籽粒,才能够满足工程化育种需要,而实际育种工作中挑选单倍体时常用的分子生物学、田间形态学辨别等方法存在耗时长、成本高、破坏样本等缺点,难以高效精准地得到玉米单倍体籽粒。相关研究已经证明高油玉米的单倍体与二倍体之间具有明显含油率差异,目前低场核磁共振技术可用于检测玉米单籽粒的含油率,并根据含油率对单倍体进行鉴别,但核磁共振仪存在价格贵、维护难、速度慢、效率低等弱点,现有设备完成单籽粒分选需用时4 s,无法满足工程化育种中大量筛选的速度需求。使用VIAVI微型近红外光谱仪能够达到0.25 s每颗的检测速度,相比核磁共振技术速度快,仪器价格较低,维护方便。使用近红外光谱仪分析技术对单倍体与二倍体籽粒进行鉴别,可以取代核磁共振鉴别单倍体的方法。采用近红外光谱定性鉴别单倍体籽粒虽然取得了一定效果,但目前研究中所采集玉米品种相对较少,研究只针对某一品种单倍体建立模型,对该品种单倍体进行分类;国内外尚无多品种混合单倍体鉴别相关研究,而工程化育种亟需一种能够识别多个品种玉米单倍体的鉴别方法。为此,本文提出一种基于深度信念网络的多品种混合玉米籽粒单倍体鉴别方法,DBN是一种多层深度神经网络,每层由受限玻尔兹曼机构成,采用逐层训练策略,可解决传统神经网络训练方法不适用于多层网络训练的问题。对比实验结果表明使用DBN方法建立多品种单倍体鉴别模型具有较高分类性能,能够满足玉米工程化育种精度要求。     

随机中毒对神经元网络时空动力学行为的影响

神经元细胞膜上的离子通道能够被一些有毒的化学物质阻断. 离子通道阻断会降低离子通道的电导率和激活通道数, 影响神经元的放电活动, 进而影响神经网络时空模式的动力学行为. 本文采用具有周期边界的近邻耦合Hodgkin-Huxley神经元网络, 数值研究了钠离子和钾离子通道随机中毒时神经网络时空模式的演化过程. 发现钠离子和钾离子通道随机中毒可以导致螺旋波破裂. 通过分析网络的放电概率, 发现钠离子通道随机中毒降低了神经网络的兴奋性, 且其对中毒的敏感程度与噪声强度有关; 钾离子通道随机中毒增强了神经网络的兴奋性. 与均匀的通道中毒相比, 随机通道中毒的神经网络具有更丰富的动力学行为. 最后, 采用无流边界条件对神经网络进行数值仿真, 得到了类似的结果. 该研究更真实地反映神经系统中毒时整体兴奋性的变化, 从另一个方面揭示离子通道中毒对网络时空行为的影响, 有利于更进一步理解离子通道在网络整体行为中的作用. 关键词： 神经网络 离子通道 随机中毒 时空动力学     

Blocking in wavelength-routed all-optical WDM network with wavelength conversion

The blocking probability in wavelength-routed all optical networks is very important measure of performance of the network, which can be affected by many factors such as network topology, traffic load, number of links, algorithms employed and whether wavelength conversion is available or not. In this paper, we have proposed a mathematical model to reduce the blocking probability of the WDM optical network for wavelength-convertible networks. The model can be used to evaluate the blocking performance of any network topology also it can be useful to improve its blocking performance of the given network topology. The blocking probability variation of the network for a particular load (per link) has been studied based on the load variation and total number of wavelengths used in the network. This model gives good results for high load (per link).