Multilayer Network Architecture Research Articles

MURE: Multi-layer real-time livestock management architecture with unmanned aerial vehicles using deep reinforcement learning

In recent years, the combination of unmanned aerial vehicles (UAVs) and wireless sensor networks (WSNs) has gained popularity in livestock management (LM) due to energy constraints and network instability. Limited energy storage of sensor nodes (SNs) and the possibility of packet loss contribute to fast energy consumption and unstable networks, respectively. UAVs serve as relay nodes and data sinks, addressing these issues by temporarily storing data to reduce SN workload and establishing mobile nodes for network stability. We propose two innovations based on previous work: 1) We introduce a multi-layer wireless network architecture, categorizing UAVs into two layers based on their functions including data collection and data processing. This enhances task parallelization, bridging performance gaps among multiple UAVs; 2) We overcome the mobility limitation of SNs, considering their real-time movement in the network. Through deep reinforcement learning, UAVs learn to cooperatively locate moving SNs. This accounts for the inevitable mobility of livestock in the industry. Additionally, we simulate the environment and compare our approach to traditional methods, evaluating metrics such as collected data per timestep (DCPS), energy consumed per timestep (ECPS), and network stability (NS). Experimental results demonstrate that our method outperforms traditional approaches, achieving a data collecting gain of 4.84% and 8.20% compared to the methods without considering SN mobility or the multi-layer characteristics of WSNs, respectively. Under energy consumption limits, our method yields energy savings of 3.00% and 1.35% respectively. Furthermore, we extensively study and validate our method against other path planning algorithms, including genetic particle swarm optimization (GPSO), modified central force optimization (MCFO), and rapidly-exploring random trees (RRT). Our approach surpasses these methods in terms of data collecting efficiency and network stability.

The CNN: The Architecture Behind Artificial Intelligence Development

The convolutional neural network (CNN) is a multilayer network architecture that is capable of training itself using an advanced algorithm to produce increasingly accurate results. The CNN is especially effective in spatial image recognition and is used in a multitude of fields, such as image recognition in the medical industry, and image segmentation in the security field. The ability of the CNN to show impressive results is seen in its multilayer composition.
 This multilayer network consists of the convolutional layer, the pooling layer, the activation layer, and the fully connected layer. Here, the convolutional layer, the pooling layer, and the activation layer have their own parameters which will influence the flow of input data throughout the CNN until it produces an ultimate output in the connected layer. As such, this paper will delve into the individual characteristics of each layer, and introduce its relationship with not only its immediate surrounding layers but with the entirety of the CNN. Additionally, this paper will stress important modules and parameters that can help improve the layers.

Mathematical modeling method based on heterogeneous cellular network algorithm and computer multi-dimensional space

Due to the rapid development of services such as computer multi-dimensional space, traditional cellular networks are quantified and diversified, and it is difficult to meet the future connection needs of green network users. The heterogeneous cellular network algorithm places multiple microcells and relay nodes under the macrocell. This multi-layer network architecture can not only shorten the data transmission link, but also reduce the energy consumption of base stations and terminals. In addition, through spectrum space reuse Capability and spectrum efficiency improve the efficiency of the system, but this also increases the number of base stations and network construction costs, and significantly increases energy consumption and operating costs of operators' electricity expenditures. In addition, in order to achieve parallel transmission of data, placing multiple millimeter wave input and output antenna systems in heterogeneous network microcells can meet the high-speed data needs of indoor users, but the antenna array requires many RF links to drive, which increases the hardware implementation. Complexity, cost and energy consumption. According to the above-mentioned problems, a reconfigurable modular robot is established based on the multi-dimensional computer space. It is composed of modules with the same interface and can be assembled into different configurations according to different tasks. Compared with the original robot, the reconfigurable modular robot is more adaptable to the work and environment, and flexible. Then the basic and core technologies such as kinematics, type synthesis, scale synthesis, error and control of the reconfigurable modular robot are mathematically modeled, so as to effectively apply and promote the reconfigurable modular robot in practice, and make it has theoretical and practical value.

Urban Road Traffic Flow Prediction with Attention-Based Convolutional Bidirectional Long Short-Term Memory Networks

Accurate traffic prediction is critical for industry practitioners and researchers in intervening and dredging future traffic in advance to avoid traffic congestion. Considering that most prediction models fail to effectively capture the complex nonlinearity of traffic data and thus cannot obtain satisfactory prediction results, we propose a novel deep-learning architecture for traffic flow prediction, called AC-BLSTM (attention-based convolutional bidirectional long short-term memory). The proposed model captures traffic information through multilayer network architectures composed of convolutional bidirectional long short-term memory (conv-BiLSTM) network and attention mechanism. The spatiotemporal features of traffic flow are extracted by convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM) network. Then attention mechanism combines the outputs of CNN and BiLSTM to assign corresponding weights to the features extracted at different times. In addition, we employ a parallel sub-module structure to model three temporal properties of traffic flow, that is, weekly, daily, and recent dependencies. Finally, the results of these three parts are fused to predict the traffic flow values through the fully connected (FC) layers. Experiment results using a real-world urban road traffic dataset demonstrate that compared with other competing models, the proposed model has better prediction performance.

Preoperative Prediction and Identification of Extracapsular Extension in Head and Neck Cancer Patients: Progress and Potential.

Background This study aimed to demonstrate both the potential and development progress in the identification of extracapsular nodal extension in head and neck cancer patients prior to surgery. Methodology A deep learning model has been developed utilizing multilayer gradient mapping-guided explainable network architecture involving a volume extractor. In addition, the gradient-weighted class activation mapping approach has been appropriated to generate a heatmap of anatomic regions indicating why the algorithm predicted extension or not. Results The prediction model shows excellent performance on the testing dataset with high values of accuracy, the area under the curve, sensitivity, and specificity of 0.926, 0.945, 0.924, and 0.930, respectively.The heatmap results show potential usefulness for some select patients but indicate the need for further training as the results may be misleading for other patients. Conclusions This work demonstrates continued progress in the identification of extracapsular nodal extension in diagnostic computed tomography prior to surgery.Continued progress stands to see the obvious potential realized where not only can unnecessary multimodality therapy be avoided but necessary therapy can be guided on a patient-specific level with information that currently is not available until postoperative pathology is complete.

Generalizing Aggregation Functions in GNNs: Building High Capacity and Robust GNNs via Nonlinear Aggregation.

The main aspect powering GNNs is the multi-layer network architecture to learn the nonlinear representation for graph learning task. The core operation in GNNs is the message propagation in which each node updates its information by aggregating the information from its neighbors. Existing GNNs usually adopt either linear neighborhood aggregation (e.g. mean, sum) or max aggregator in their message propagation. 1) For linear aggregators, the whole nonlinearity and network's capacity of GNNs are generally limited because deeper GNNs usually suffer from the over-smoothing issue due to their inherent information propagation mechanism. Also, linear aggregators are usually vulnerable to the spatial perturbations. 2) For max aggregator, it usually fails to be aware of the detailed information of node representations within neighborhood. To overcome these issues, we re-think the message propagation mechanism in GNNs and develop the new general nonlinear aggregators for neighborhood information aggregation in GNNs. One main aspect of our nonlinear aggregators is that they all provide the optimally balanced aggregator between max and mean/sum aggregators. Thus, they can inherit both i) high nonlinearity that enhances network's capacity, robustness and ii) detail-sensitivity that is aware of the detailed information of node representations in GNNs' message propagation. Promising experiments show the effectiveness, high capacity and robustness of the proposed methods.

A Dynamic Handover Software-Defined Transmission Control Scheme in Space-Air-Ground Integrated Networks

The Space-Air-Ground Integrated Networks (SAGINs) converges the rapidly developing ground, aerial and satellite communication networks to provide users with efficient and personalized services through the multi-layer network architecture. The complex structure, inconsistent communication protocol, and incompatible equipment significantly affect the information transmission efficiency in SAGINs. The application of Software Defined Network (SDN) technology promotes global deployment of the SAGIN resources. Aiming at the information transmission bottleneck between controller and switch nodes in the software-defined SAGINs, a dynamic handover transmission control scheme based on the queuing game model is proposed. Considering long transmission delay links in the SAGINs, the traditional queuing game model is improved to reduce the number of information interactions. The users’ service value are defined as the random distribution to describe the various requirements in the network. In addition, the relationship between the social welfare and the arrival rate of the transmission control system is discussed under two modes of observable and unobservable controller queues. Through theoretical analysis and calculation, the unique handover arrival rate is obtained to make the social welfare equal in the two modes. When the arrival rate is less than the handover threshold, the controller queue is set to unobservable mode, and otherwise to observable mode. Then, the system can operate in an optimal way to obtain the maximum benefit. At last, numerical and system simulation verify the effectiveness of the scheme.

Network Audio Data and Music Composition Teaching Based on Heterogeneous Cellular Network.

With the rapid development of services such as Industry 4.0 and Internet of Vehicles, it is difficult for traditional cellular networks to meet the needs of network users for quantification, diversification, and greenness in the future. Various cellular networks expand multiple micro-cell nodes and relay nodes under macro-cells to form a multilayer network architecture. Based on this, in the process of data transmission, the links have been repeatedly reduced, and at the same time, the terminal power consumption has been reduced and the running system has been improved. This article will use the ratio of the capacity, energy consumption, and resource allocation of different cellular networks as the main means to optimize the cost. Using graph theory, auction theory, and multipurpose optimization algorithms, we have conducted in-depth research topics on upstream and downstream wireless resource allocation, network relay deployment and transmission scheduling, MMW large-scale multi-antenna transmission technology, and base station energy management. A series of optimization schemes and algorithms are proposed. This dissertation is based on the research of educational system design theory in the field of educational technology so as to carry out the research of music education system design theory suitable for the nature of music subjects and learning and education characteristics. Based on the necessity and importance of music education system design theory, the research framework of music education system design theory is constructed in advance. The voice data acquisition system collects voice data through a network grabber and real-time recording and uses signal processing and pattern recognition technology to automatically classify the collected voice data into three categories: voice, environmental sound, and music. After establishing the audio data deployment strategy, simulation method, and architecture design based on heterogeneous cellular network, this paper designs the corresponding music composition teaching system, mainly including score editing, viewing, and content display of the composition teaching system, and the final test shows that the system designed in this paper can be effectively used in various music school teaching combined with heterogeneous cellular networks.

Multilabel Text Classification Using Multilayer DGAT

Text classification is one of the most fundamental tasks in text data analysis. Generally, textual data are extensive in scale and complex in the inherent relationship, leading to low accuracy in traditional classification models. This paper proposes a multi-label classification model based on multi-layer neural network architecture. We first construct a dual-attention mechanism graph neural network (named DGAT for short) to fuse the typological and informational features of the target node and the connected nodes. Secondly, we also build a multi-layer network architecture with multiple DGATs to expand the range of neighborhood nodes participating in the feature fusion to meet the needs of classifying different datasets. To ensure the learning ability of the model, we also use the residual network to solve the problem of error rise and gradient descent caused by multi-layer network architecture. Finally, we conducted a large number of experiments on five benchmark datasets. The results show that the accuracy of the proposed model is significantly better than that of the traditional models, and that there is a noticeable improvement when compared with other deep learning methods.

Employing FAMTAR and AHB to Achieve an Optical Resource-Efficient Multilayer IP-Over-EON SDN Network

With the global Internet traffic continuously growing, network operators face more and more challenges related to the management of their networks. Efficient utilization of the available network resources becomes crucial to maintain the desired Quality of Service level and control the upsurge in operational expenses. The combination of the Software-Defined Networking concept with the multi-layer network architecture can simplify the process of control and management of the network, its layers, and resources. In this paper, we propose a solution to enhance resource utilization in software-defined multi-layer optical networks. The proposed solution takes the advantages of the Software-Defined Networking, Flow-Aware Multi-Topology Adaptive Routing, and Automatic Hidden Bypasses mechanisms to ensure simultaneous, multi-path data transmissions in both IP and optical layers. The Software-Defined Networking controller manages both mentioned mechanisms, selects the best possible bypass, and allocates lightpaths to ensure that the optical spectral efficiency is optimal. The evaluation shows that the proposed solution for multi-layer software-defined network increases the overall network performance and resource utilization.

Machine Learning Classifiers for Surface Crack Detection in Fracture Experiments

Correctly determining the onset of fracture is crucial when performing mechanical experiments. Commonly carried out by visual inspection, here an image-based machine learning approach is proposed to classify cracked and un-cracked specimens. It yields the potential to objectify and automate crack detection, thereby removing sources of uncertainty and error from the post-processing of experiments. More than 30’000 speckle-pattern images obtained from 77 experiments on three specimen geometries are evaluated. They comprise uniaxial tension, notched tension as well as axisymmetric V-bending experiments. Statistical texture features are extracted from all images. They include both first-order (variance, skewness, kurtosis) and higher-order statistical texture features, i.e. Haralick features. The discriminatory power of the texture information is evaluated based on the Fisher's Discriminant Ratio and feature correlations are identified and quantified. Image texture feature subsets of high discriminatory power are used to parse neural network architectures of different complexities from simple perceptron to feed-forward and cascade neural networks. It is found that a small subset of the investigated texture features is highly significant for all experiments. Using this feature subset in conjunction with multi-layer, non-linear and low complexity feed-forward network architectures classification accuracies in the order of 99% are obtained. At the same time, it is shown that linear classifiers are not sufficient to robustly distinguish the state of the specimens, even when high discriminatory power features are used.Graphical Abstract: [Display omitted]

A Sparsity-Driven Backpropagation-Less Learning Framework Using Populations of Spiking Growth Transform Neurons.

Growth-transform (GT) neurons and their population models allow for independent control over the spiking statistics and the transient population dynamics while optimizing a physically plausible distributed energy functional involving continuous-valued neural variables. In this paper we describe a backpropagation-less learning approach to train a network of spiking GT neurons by enforcing sparsity constraints on the overall network spiking activity. The key features of the model and the proposed learning framework are: (a) spike responses are generated as a result of constraint violation and hence can be viewed as Lagrangian parameters; (b) the optimal parameters for a given task can be learned using neurally relevant local learning rules and in an online manner; (c) the network optimizes itself to encode the solution with as few spikes as possible (sparsity); (d) the network optimizes itself to operate at a solution with the maximum dynamic range and away from saturation; and (e) the framework is flexible enough to incorporate additional structural and connectivity constraints on the network. As a result, the proposed formulation is attractive for designing neuromorphic tinyML systems that are constrained in energy, resources, and network structure. In this paper, we show how the approach could be used for unsupervised and supervised learning such that minimizing a training error is equivalent to minimizing the overall spiking activity across the network. We then build on this framework to implement three different multi-layer spiking network architectures with progressively increasing flexibility in training and consequently, sparsity. We demonstrate the applicability of the proposed algorithm for resource-efficient learning using a publicly available machine olfaction dataset with unique challenges like sensor drift and a wide range of stimulus concentrations. In all of these case studies we show that a GT network trained using the proposed learning approach is able to minimize the network-level spiking activity while producing classification accuracy that are comparable to standard approaches on the same dataset.

An efficient Time-sensitive data scheduling approach for Wireless Sensor Networks in smart cities

The Continuous increase in urban population causes enormous pressure on cities’ limited resources, including transport, energy, water, housing, public services, and others. Hence, the need to plan and develop smart cities based solutions for enhanced urban governance is becoming more evident. These solutions are motivated by innovations in Information and Communication Technology to support smart planning for the city and to facilitate enhanced services to its citizens. Important areas where smart city services can be offered include urban planning, transport planning, energy conservation, water management, waste management, environmental monitoring, public safety, healthcare, education, entertainment, and many other services. Hence, the enormous data collected from different networks and applications to facilitate the offering of smart city services requires efficient data scheduling, aggregation, and processing to ensure service quality (QoS). However, existing data scheduling approaches consider scheduling and processing data only in the cloud, while processing also in the data collecting devices is significantly essential. This paper first introduces the multi-layer network architecture comprising sensor/device networks and cloud. The paper then introduces a Multi-layer, Priority-based, Dynamic, and Time-sensitive data processing and Scheduling approach (MPDTS) in the proposed multi-layer networks. Simulation results show that the proposed MPDTS approach achieves lower latency and data processing time than existing traditional data scheduling approaches that work only in the cloud layer.

Klasifikasi American Sign Language Menggunakan Ekstraksi Fitur Histogram of Oriented Gradients dan Jaringan Syaraf Tiruan

Sign languages have various types, one of which is American Sign Language (ASL). In this study, ASL images from the handshape alphabet were extracted using Histogram of Oriented Gradient (HOG) then these features were used for the classification of Artificial Neural Networks (ANN) with various training functions using 3 variations of multi-layer network architecture where ANN architecture consists of one hidden layer. Based on ANN training, trainbr test results have a higher success rate than other training functions. In architecture with 15 neurons in the hidden layer get an accuracy value of 99.29%, a precision of 91.84%, and a recall of 91.47%. The test results show that using the HOG feature and ANN classification method for ASL recognition gives a good level of accuracy, with an overall accuracy of 5 neurons 95.38%, 10 neurons 96.64%, and 15 neurons with 97.32%.
 
 Keywords— Artificial Neural Network; American Sign Language; Histogram of Oriented Gradient; Training Function

Multi-Layer Network Security Architecture

This month is the last of four installments on multi-layer network architectures. We talked about the need for satellite, fiber and wireless network architectures to include all layers in its designs. When they are integrated together to form one Internet, issues on internetworking appear at all layers including the Application Layer. All the more critical are network architectures with coherent multi-layer designs. This last piece addresses network security and resiliency and the need for a multi-layer approach to an effective design.

Integration of Visible Light Communication and Positioning within 5G Networks for Internet of Things

With the widespread deployment of Internet of Things (IoT), more and more devices are involved in wireless networks, and the fifth generation (5G) network requires to support the massive connectivity and diverse services for the huge number of IoT devices. Visible light communications (VLC) and visible light positioning (VLP) are two promising supplementary technologies to assist 5G networks to support the massive connectivity, high reliability, high data rate, high positioning accuracy, low latency, low power consumption and improved security of IoT. Hence, this article presents a multi-layer network architecture by integrating VLC and VLP within 5G networks, in order to support the above mentioned diverse requirements of IoT devices. In the multi-layer network, the macrocell and picocell layers support better coverage and reliability via the radio frequency (RF) spectrum, while the optical attocell layer provides the high-speed transmission and high-accuracy positioning services operating at the visible light spectrum. We then briefly describe some key technologies for the performance improvement of the multi-layer network, including energy harvesting, modulation and multiple access schemes. An exemplary case study and simulation analysis are provided to demonstrate the advantage and significance of the presented multi-layer network for IoT. Finally, we point out some future research directions.

Software-Defined Space-Air-Ground Integrated Network Architecture with the Multi-Layer Satellite Backbone Network

Under the background of the rapid development of ground mobile communication, the advantages of high coverage, survivability, and flexibility of satellite communication provide air support to the construction of space information network. According to the requirements of the future space information communication, a software-defined Space-Air-Ground Integrated network architecture was proposed. It consisted of layered structure satellite backbone network, deep space communication network, the stratosphere communication network and the ground network. The SpaceAir-Ground Integrated network was supported by the satellite backbone network. It provided data relay for the missions such as deep space exploration and controlled the deep-space spacecraft when needed. In addition, it safeguarded the anti-destructibility of stratospheric communication and assisted the stratosphere to supplement ground network communication. In this paper, algorithm requirements of the congestion control and routing of satellite backbone protocols for heterogeneous users’ services were proposed. The algorithm requirements of distinguishing different service objects for the deep space communication and stratospheric communication network protocols were described. Considering the realistic demand for the dynamic coverage of the satellite backbone network and node cost, the multi-layer satellite backbone network architecture was constructed. On this basis, the proposed Software-defined Space-Air-Ground Integrated network architecture could be built as a large, scalable and efficient communication network that could be integrated into space, air, and ground.

PRACTICAL IMPLEMENTATION OF THE SOFTWARE DEFINED 5G MOBILE NETWORK BASED ON CLOUD-RAN AND SDR TECHNOLOGIES

The paper is devoted to the practical implementation aspects of a software-defined mobile network based on the Cloud-RAN architecture and universal software radio peripheral National Instruments USRP 2900. We propose a multilayer network architecture, which consists of a radio access network (RAN) plane, a core plane, a control plane, an artificial intelligence (AI) plane, and a monitoring system that collects data about network performance. The RAN plane provides all functions related to channel scheduling, data encoding and signal processing and combines all macro and small cells, as well as Wi-Fi access points. The core plane provides functionality of the evolved packet core such as service and packet gateways, mobility management entities and edge routers. The control plane is responsible for load balancing over all cells, spectrum sharing, spectrum reallocation, infrastructure reconfiguration, and other management parameters. The AI plane is responsible for inference and decisions making about optimal network configuration taking into account instantaneous network conditions. Such decisions are supported by a set of machine learning models, which are trained over the available statistics, collected through a long-term period of network operation. External monitoring system provides the needed data for AI plane by using the MQTT (Message Queue Telemetry Transport) protocol to collect information from all base stations, Wi-Fi access points and mobile devices.The proposed system enables the logical network slicing into multiple isolated segments over a single physical infrastructure by using virtualization of radio resources. A new method of intelligent radio resource management is proposed to improve the efficiency of spectrum utilization by predicting traffic demand of particular network sliced using deep recurrent neural network. Simulation results show that for the case of network slicing the proposed method allows to double the capacity of the physical network infrastructure, while ensuring the target quality of service for all users.

Forecast Based Handover in an Extensible Multi-Layer LEO Mobile Satellite System

Low earth orbit mobile satellite system (LEO-MSS) is the major system to provide communication support for the regions beyond the coverage of terrestrial network systems. However, passive handover happens frequently caused by the quick movement of LEO satellites in LEO-MSS. It not only causes the waste of radio resource, but also makes it hard to guarantee the quality of service (QoS), especially for user groups in hot-spot regions. To tackle this problem, we propose an extensible multi-layer network architecture to reduce the handover rate, especially group handover rate by introducing high-altitude platforms (HAPs) and terrestrial relays (TRs) to this system. We then propose a multi-layer handover management framework and also design different handover procedures based on handover forecast for different kinds of handovers according to the proposed architecture and framework to reduce handover delay and signalling cost. Furthermore, we propose a dynamic handover optimization to reduce the dropping probability and guarantee the QoS of mobile terminals. Numerical results show that the proposed architecture reduces group handovers significantly. The proposed handover procedures also provide better performance on delay and signalling cost compared with traditional handover protocols. With the proposed dynamic handover optimization, the proposed handover procedures provide better performance on dropping probability and throughput. The proposed dynamic handover optimization has an excellent performance on dropping probability while guaranteeing the QoS of mobile terminals.

A survey on multi-layer IP and optical Software-Defined Networks

As Software-Defined Networks become more and more popular, they start to appear in various architectures. The most common utilization of SDN is to control the network in the electric layers, including OSI/ISO layers 2 and above. However, SDN can be used much further. Professionals and scientists noticed that combining management and control of many layers at the same time can provide benefits. At one point, SDN started to be envisioned for multi-layer networks.In this survey we present, compare and contrast solutions that utilize SDN in multi-layer network architectures. The main objective is to analyze evolving multi-layer network architectures and show how these solutions coupled with SDN contribute to make future networks simple, flexible and cost-effective.