基于LeNet-5的手写数字识别的改进方法

手写体识别是计算机视觉的一个重要研究方向。在手写体识别中,常规方法的泛化性能通常较低。相比之下,人工神经网络能够从样本数据中学习特征表达。文章详细探讨了基于LeNet-5和基于卷积神经网络的手写数字识别方法,并设计了图形用户界面(GUI)进行实际测试。测试结果显示,改进后的LeNet-5模型在手写数字识别上相较于传统LeNet-5模型有一定提升。

基于权重分摊的LeNet-5卷积神经网络防御策略

随着神经网络在自动驾驶、医疗诊断等关键领域的应用不断深入,如何确保神经网络的鲁棒性和安全性已成为当前研究的热点和挑战。在对抗攻击、数据中毒攻击、后门攻击等众多攻击方式中,随机翻转攻击是一种对安全性影响极大的攻击,其通过改变模型内部的权重参数来攻击网络,以降低网络性能。为应对此攻击方式,研究了一种基于权重分摊的防御策略。通过计算和分析权重的梯度来确定关键神经元,并为这些神经元添加冗余结构,使错误的权重最终被稀释,以提高模型的容错能力。为了验证这一防御策略,以LeNet-5模型为实验对象进行实验。实验表明,在相同的攻击条件下,经过防御后的模型相较于原始LeNet-5模型,容错精度提升了6.5%,相较于Inception-LeNet-5模型在全连接层上容错精度提升了1.9%。

基于改进LeNet-5网络的堆芯燃料组件编码识别

在核电站堆芯核燃料组件水下组装作业中,需要通过视觉技术进行组件编码的识别以便准确定位组件的安装位置。针对水下环境中弱光照等问题导致了图像质量的降低,本文通过乘方增强算法、OSTU算法、CLAHE算法和拉普拉斯变换的方法来实现堆芯燃料组件编码字符水下图像的增强。为了提高编码识别效果,提出了一种整合LeNet-5网络和支持向量机(SVM)的模型,在网络中添加BN(Batch Normalization)层与Dropout层来加速网络的运行速度,并改进Sigmoid函数,增加函数的平滑性,以此来减少梯度消失。实验表明,在自定义数据集上的验证准确率为99.82%,识别率为100%,相比于其他模型有显著的提升。

Channel assignment and power allocation for throughput improvement with PPO in B5G heterogeneous edge networks

In Beyond the Fifth Generation(B5G)heterogeneous edge networks,numerous users are multiplexed on a channel or served on the same frequency resource block,in which case the transmitter applies coding and the receiver uses interference cancellation.Unfortunately,uncoordinated radio resource allocation can reduce system throughput and lead to user inequity,for this reason,in this paper,channel allocation and power allocation problems are formulated to maximize the system sum rate and minimum user achievable rate.Since the construction model is non-convex and the response variables are high-dimensional,a distributed Deep Reinforcement Learning(DRL)framework called distributed Proximal Policy Optimization(PPO)is proposed to allocate or assign resources.Specifically,several simulated agents are trained in a heterogeneous environment to find robust behaviors that perform well in channel assignment and power allocation.Moreover,agents in the collection stage slow down,which hinders the learning of other agents.Therefore,a preemption strategy is further proposed in this paper to optimize the distributed PPO,form DP-PPO and successfully mitigate the straggler problem.The experimental results show that our mechanism named DP-PPO improves the performance over other DRL methods.

Artificial Neural Network and Fuzzy Logic Based Techniques for Numerical Modeling and Prediction of Aluminum-5%Magnesium Alloy Doped with REM Neodymium

In this study, the mechanical properties of aluminum-5%magnesium doped with rare earth metal neodymium were evaluated. Fuzzy logic (FL) and artificial neural network (ANN) were used to model the mechanical properties of aluminum-5%magnesium (0-0.9 wt%) neodymium. The single input (SI) to the fuzzy logic and artificial neural network models was the percentage weight of neodymium, while the multiple outputs (MO) were average grain size, ultimate tensile strength, yield strength elongation and hardness. The fuzzy logic-based model showed more accurate prediction than the artificial neutral network-based model in terms of the correlation coefficient values (R).

基于1D-LeNet-5模型的滚动轴承故障诊断方法

风力发电过程中,轴承能否正常运行关系到风电机组能否正常工作.针对现有基于深度学习的轴承故障诊断模型结构复杂、参数众多和训练困难的问题,提出了基于LeNet-5模型改进的一维卷积神经网络滚动轴承故障诊断方法.首先,为了更大程度提取故障信息,引入短时傅里叶变换对原始振动信号进行预处理.其次,设计一维网络模型,其感受野更大,计算速度更快;同时,引入Leaky-ReLU激活函数,其对输入信号的细节处理能力更强;并且增加批归一化层和Dropout层,提高模型泛化能力.最后,利用训练后的模型进行故障诊断实验.结果表明,该方法在10类轴承故障分类中诊断准确率能够达到99.98%,针对风电机组轴承故障诊断具有较好的工程应用前景.

基于字符分割和LeNet-5网络的字符验证码识别

为了解决传统验证码识别方法效率低,精度差的问题,设计了一种先分割后识别的验证码处理方案;该方案在预处理阶段用中值滤波去噪,再利用霍夫变换对图像字符进行矫正;在字符分割阶段,利用垂直投影算法确定验证码字符块个数,以及字符坐标点,再用颜色填充算法对验证码进行初步分割,根据分割后的字符块数量对粘连字符进行二次分割;在识别阶段,我们对LeNet-5网络进行了改进,修改了输入层,并用全连接层替换了LeNet-5网络中的C5层,以此来对验证码字符进行识别;实验表明,对于非粘连验证码和粘连验证码,单张图片分割时间为0.14和0.15 ms,分割准确率为98.75%和97.25%,识别准确率为99.99%和97.7%;结果表明,该算法对验证码分割和识别都有着很好的效果。

基于改进LeNet-5优化算法的轴承故障诊断研究

针对直升机自动倾斜器滚动轴承振动信号复杂而传统卷积神经网络对轴承故障信号微小特征提取困难导致的故障诊断精度不高的问题,提出基于LeNet-5网络的一种改进方法。首先,在LeNet-5网络中设计一个新的特征提取模块,形成并行的特征提取框架,增强网络对微小特征的提取能力,缓解直升机故障诊断精度不高的问题。其次,采用Dropout层和自适应的参数算法,避免模型不稳定,加速模型的收敛。最后,利用课题组轴承数据和西储大学公开数据集开展实验,结果表明,相较于原LeNet-5网络模型,改进后的LeNet-5网络具有较高的测试精度,在课题组数据集的测试精度达99.6%,西储大学数据集的测试精度为100%,说明该模型对滚动轴承的故障诊断具有更高的准确率。

Machine Learning Based Classifiers for QoE Prediction Framework in Video Streaming over 5G Wireless Networks

Recently,the combination of video services and 5G networks have been gaining attention in the wireless communication realm.With the brisk advancement in 5G network usage and the massive popularity of threedimensional video streaming,the quality of experience(QoE)of video in 5G systems has been receiving overwhelming significance from both customers and service provider ends.Therefore,effectively categorizing QoE-aware video streaming is imperative for achieving greater client satisfaction.This work makes the following contribution:First,a simulation platform based on NS-3 is introduced to analyze and improve the performance of video services.The simulation is formulated to offer real-time measurements,saving the expensive expenses associated with real-world equipment.Second,A valuable framework for QoE-aware video streaming categorization is introduced in 5G networks based on machine learning(ML)by incorporating the hyperparameter tuning(HPT)principle.It implements an enhanced hyperparameter tuning(EHPT)ensemble and decision tree(DT)classifier for video streaming categorization.The performance of the ML approach is assessed by considering precision,accuracy,recall,and computation time metrics for manifesting the superiority of these classifiers regarding video streaming categorization.This paper demonstrates that our ML classifiers achieve QoE prediction accuracy of 92.59%for(EHPT)ensemble and 87.037%for decision tree(DT)classifiers.

Dynamic QoS Mapping and Adaptive Semi-Persistent Scheduling in 5G-TSN Integrated Networks

The ubiquitous and deterministic communication systems are becoming indispensable for future vertical applications such as industrial automation systems and smart grids.5G-TSN(Time-Sensitive Networking)integrated networks with the 5G system(5GS)as a TSN bridge are promising to provide the required communication service.To guarantee the endto-end(E2E)QoS(Quality of Service)performance of traffic is a great challenge in 5G-TSN integrated networks.A dynamic QoS mapping method is proposed in this paper.It is based on the improved K-means clustering algorithm and the rough set theory(IKCRQM).The IKC-RQM designs a dynamic and loadaware QoS mapping algorithm to improve its flexibility.An adaptive semi-persistent scheduling(ASPS)mechanism is proposed to solve the challenging deterministic scheduling in 5GS.It includes two parts:one part is the persistent resource allocation for timesensitive flows,and the other part is the dynamic resource allocation based on the max-min fair share algorithm.Simulation results show that the proposed IKC-RQM algorithm achieves flexible and appropriate QoS mapping,and the ASPS performs corresponding resource allocations to guarantee the deterministic transmissions of time-sensitive flows in 5G-TSN integrated networks.

From Earth to Space:A First Deployment of 5G Core Network on Satellite

Recent developments in the aerospace industry have led to a dramatic reduction in the manufacturing and launch costs of low Earth orbit satellites.The new trend enables the paradigm shift of satelliteterrestrial integrated networks with global coverage.In particular,the integration of 5G communication systems and satellites has the potential to restructure nextgeneration mobile networks.By leveraging the network function virtualization and network slicing,the satellite 5G core networks will facilitate the coordination and management of network functions in satellite-terrestrial integrated networks.We are the first to deploy a 5G core network on a real-world satellite to investigate its feasibility.We conducted experiments to validate the satellite 5G core network functions.The validated procedures include registration and session setup procedures.The results show that the satellite 5G core network can function normally and generate correct signaling.

基于改进LeNet-5网络的数字电表识别

目前在水下等特殊应用场景的电表识别研究中,虽然LeNet-5网络表现良好,但仍存在泛化能力不足、鲁棒性较差等问题。为此,文章基于改进LeNet-5网络的数字电表识别方法,通过增加激活离群值去除,利用dropout算法和ReLU激活函数增强神经网络泛化能力与鲁棒性。实验结果表明:改进的LeNet-5网络模型在学习速率为0.1%和迭代次数为600次时,网络精度达到99.42%。该方法具有较强的运算能力和较高的网络识别精度,可满足水下数字电表识别需求。

Optical Neural Networks:Analysis and Prospects for 5G Applications

With the capacities of self-learning,acquainted capacities,high-speed looking for ideal arrangements,solid nonlin-ear fitting,and mapping self-assertively complex nonlinear relations,neural systems have made incredible advances and accomplished broad application over the final half-century.As one of the foremost conspicuous methods for fake insights,neural systems are growing toward high computational speed and moo control utilization.Due to the inborn impediments of electronic gadgets,it may be troublesome for electronic-implemented neural systems to make the strides these two exhibitions encourage.Optical neural systems can combine optoelectronic procedures and neural organization models to provide ways to break the bottleneck.This paper outlines optical neural networks of feedforward repetitive and spiking models to give a clearer picture of history,wildernesses,and future optical neural systems.The framework demonstrates neural systems in optic communication with the serial and parallel setup.The graphene-based laser structure for fiber optic communication is discussed.The comparison of different balance plans for photonic neural systems is made within the setting of hereditary calculation and molecule swarm optimization.In expansion,the execution comparison of routine photonic neural,time-domain with and without extending commotion is additionally expounded.The challenges and future patterns of optical neural systems on the growing scale and applications of in situ preparing nonlinear computing will hence be uncovered.

PSAP-WSN:A Provably Secure Authentication Protocol for 5G-Based Wireless Sensor Networks

Nowadays,the widespread application of 5G has promoted rapid development in different areas,particularly in the Internet of Things(IoT),where 5G provides the advantages of higher data transfer rate,lower latency,and widespread connections.Wireless sensor networks(WSNs),which comprise various sensors,are crucial components of IoT.The main functions of WSN include providing users with real-time monitoring information,deploying regional information collection,and synchronizing with the Internet.Security in WSNs is becoming increasingly essential because of the across-the-board nature of wireless technology in many fields.Recently,Yu et al.proposed a user authentication protocol forWSN.However,their design is vulnerable to sensor capture and temporary information disclosure attacks.Thus,in this study,an improved protocol called PSAP-WSNis proposed.The security of PSAP-WSN is demonstrated by employing the ROR model,BAN logic,and ProVerif tool for the analysis.The experimental evaluation shows that our design is more efficient and suitable forWSN environments.

Cooperative Relay Networks Based on the OAM Technique for 5G Applications

Orbital Angular Momentum(OAM)is an intrinsic property of electro-magnetic waves.Great research has been witnessed in the last decades aiming at exploiting the OAM wave property in different areas in radio and optics.One pro-mising area of particular interest is to enhance the efficiency of the available communications spectrum.However,adopting OAM-based solutions is not priceless as these suffer from wave divergence especially when the OAM order is high.This shall limit the practical communications distance,especially in the radio regime.In this paper,we propose a cooperative OAM relaying system consisting of a source,relay,and destination.Relays help the source to transmit packets to the destination by providing an alternative connection between source and desti-nation.This cooperative solution aims on the one hand,through best-path selection,on increasing the communications range.On the other hand,through the parallel transmission orders allowed by OAM carrying waves,the system could raise its total transmission throughput.Simulation results show that combining a cooperative relay with OAM improves the system throughput compared to using each element separately.In addition,the proposed cooperative relaying OAM out-performs the cooperative relaying non-orthogonal multiple access scheme,which is a key spectrally efficient technique used in 5G technology.

An Intelligent Admission Control Scheme for Dynamic Slice Handover Policy in 5G Network Slicing

5G use cases,for example enhanced mobile broadband(eMBB),massive machine-type communications(mMTC),and an ultra-reliable low latency communication(URLLC),need a network architecture capable of sustaining stringent latency and bandwidth requirements;thus,it should be extremely flexible and dynamic.Slicing enables service providers to develop various network slice architectures.As users travel from one coverage region to another area,the callmust be routed to a slice thatmeets the same or different expectations.This research aims to develop and evaluate an algorithm to make handover decisions appearing in 5G sliced networks.Rules of thumb which indicates the accuracy regarding the training data classification schemes within machine learning should be considered for validation and selection of the appropriate machine learning strategies.Therefore,this study discusses the network model’s design and implementation of self-optimization Fuzzy Qlearning of the decision-making algorithm for slice handover.The algorithm’s performance is assessed by means of connection-level metrics considering the Quality of Service(QoS),specifically the probability of the new call to be blocked and the probability of a handoff call being dropped.Hence,within the network model,the call admission control(AC)method is modeled by leveraging supervised learning algorithm as prior knowledge of additional capacity.Moreover,to mitigate high complexity,the integration of fuzzy logic as well as Fuzzy Q-Learning is used to discretize state and the corresponding action spaces.The results generated from our proposal surpass the traditional methods without the use of supervised learning and fuzzy-Q learning.

ML-SLD:A message-level stateless design for cloud-native 5G core network

The Internet of Things(IoTs)has become an essential component of the 5th Generation(5G)network and beyond,accelerating the transition to digital society.The increasing signaling traffic generated by billions of IoT devices has placed significant strain on the 5G Core network(5GC)control plane.To address this issue,the 3rd Gener-ation Partnership Project(3GPP)first proposed a Service-Based Architecture(SBA),intending to create a flexible,scalable,and agile cloud-native 5GC.However,considering the coupling of protocol states and functions,there are still many challenges to fully utilize the benefits of the cloud computing and orchestrate the 5GC in a cloud-native manner.We propose a Message-Level StateLess Design(ML-SLD)to provide a cloud-native 5GC from an architectural standpoint in this paper.Firstly,we propose an innovative mechanism for servitization of the N2 interface to maintain the connection between Radio Access Network(RAN)and the 5GC,avoiding interruptions and dropouts of large-scale user data.Furthermore,we propose an On-demand Message Forwarding(OMF)al-gorithm to reduce the impact of cloud fluctuations on the performance of cloud-native 5GC.Finally,we create a prototype that is based on the OpenAirInterface(OAI)5G core network projects,with all Network Functions(NFs)packaged in dockers and deployed in a kubernetes-based cloud environment.Several experiments have been built with UERANSIM and Chaosblade simulation tools.The findings demonstrate the viability and efficiency of our proposed methods.

An Optimized Approach for Spectrum Utilization in mmWave Massive MIMO 5G Wireless Networks

Massive multiple-input multiple-output(MIMO)systems that use the millimeter-wave(mm-wave)band have a higher frequency and more antennas,which leads to significant path loss,high power consumption,and server interference.Due to these issues,the spectrum efficiency is significantly reduced,making spectral efficiency improvement an important research topic for 5G communication.Together with communication in the terahertz(THz)bands,mmWave communication is currently a component of the 5G standards and is seen as a solution to the commercial bandwidth shortage.The quantity of continuous,mostly untapped bandwidth in the 30–300 GHz band has presented a rare opportunity to boost the capacity of wireless networks.The wireless communications and consumer electronics industries have recently paid a lot of attention to wireless data transfer and media streaming in the mmWave frequency range.Simple massive MIMO beamforming technology cannot successfully prevent interference between multiple networks in the current spectrum-sharing schemes,particularly the complex interference dispersed in indoor communication systems such as homes,workplaces,and stadiums.To effectively improve spectrum utilization and reduce co-channel interference,this paper proposes a novel algorithm.The main idea is to utilize the spectrum in software-defined mmWave massive MIMO networks through coordinated and unified management.Then,the optimal interference threshold is determined through the beam alignment method.Finally,a greedy optimization algorithm is used to allocate optimal spectral resources to the users.Simulation results show that the proposed algorithm improved spectral efficiency and reduced interference.

Anti-jamming channel access in 5G ultra-dense networks: a game-theoretic learning approach

This paper investigates the Quality of Experience(QoE)oriented channel access anti-jamming problem in 5th Generation Mobile Communication(5G)ultra-dense networks.Firstly,considering that the 5G base station adopts beamforming technology,an anti-jamming model under Space Division Multiple Access(SDMA)conditions is proposed.Secondly,the confrontational relationship between users and the jammer is formulated as a Stackelberg game.Besides,to achieve global optimization,we design a local cooperation mechanism for users and formulate the cooperation and competition among users as a local altruistic game.By proving that the local altruistic game is an Exact Potential Game(EPG),we further prove the existence of pure strategy Nash Equilibrium(NE)among users and Stackelberg Equilibrium(SE)between users and jammer.Thirdly,to obtain the equilibrium solutions of the proposed games,we propose an anti-jamming channel selection algorithm and improve its convergence speed through heterogeneous learning parameters.The simulation results validate the convergence and effectiveness of the proposed algorithm.Compared with the throughput optimization scheme,our proposed scheme obtain a greater network satisfaction rate.Finally,we also analyze user fairness changes during the algorithm convergence process and get some interesting conclusions.

Fault Diagnosis of 5G Networks Based on Digital Twin Model

Fault diagnosis of 5G networks faces the challenges of heavy reliance on human experience and insufficient fault samples and relevant monitoring data.The digital twin technology can realize the interaction between virtual space and physical space through the fusion of model and data,providing a new paradigm for fault diagnosis.In this paper,we first propose a network digital twin model and apply it to 5G network diagnosis.We then use an improved Average Wasserstein GAN with Gradient Penalty(AWGAN-GP)method to discover and predict failures in the twin network.Finally,we use XGBoost algorithm to locate the faults in physical network in real time.Extensive simulation results show that the proposed approach can significantly increase fault prediction and diagnosis accuracy in the case of a small number of labeled failure samples in 5G networks.