Network Performance Research Articles

Multi-Type Ship Target Detection in Complex Marine Background Based on YOLOv11

Realizing accurate control of ship target information in complex marine environments is of great significance for maintaining marine environment security and safeguarding maritime sovereignty. With the rapid development of material technology and manufacturing industry, the types and styles of ships are increasing, and the distribution of multi-type ships on the sea is widespread. How to realize the accurate detection and identification of dynamic multi-type ship targets in the complex marine environment is an important and difficult problem that needs to be solved urgently in current marine environment detection. In this paper, an improved YOLOv11 ship target detection algorithm is proposed, which firstly utilizes the improved EfficientNetv2 network to replace the original backbone network of YOLOv11 to improve the learning ability of ship features under complex sea conditions; in order to solve the problem of interference by moving objects at sea when detecting dense ship targets and reduce the problems of missing detection and false alarms, the algorithm borrows from ConvNext block idea in the process of a neck feature pyramid network fusion; the algorithm introduces the WIoU loss function, which compensates for the effect of the small number of pixels of the small target in the process of regression loss computation, so as to improve the network’s performance in detecting small targets. In order to test the network performance in actual application scenarios, the article builds a visible ship target dataset, including complex background, occlusion and overlap, small targets, and other factors. Through experimental verification, the detection accuracy of the improved algorithm is improved by 5.6% compared with the original algorithm, and compared with typical algorithms in terms of detection accuracy, speed, and number of parameters, ablation experiments are designed to comprehensively validate and analyze the algorithm’s performance.

An Intelligent Fuzzy-Based Routing Protocol for Vehicular Opportunistic Networks

Opportunistic networks are characterized by intermittent connectivity and dynamic topologies, which pose significant challenges for efficient message delivery, resource management, and routing decision-making. This paper introduces the Fuzzy Control Routing Protocol, a novel approach designed to address these challenges by leveraging fuzzy logic to enhance routing decisions and improve overall network performance. The protocol considers buffer occupancy, angle to destination, and the number of unique connections of the target nodes to make context-aware routing decisions. It was implemented and evaluated using the FuzzyC framework for simulations and the opportunistic network environment simulator for realistic network scenarios. Simulation results demonstrate that the Fuzzy Control Routing Protocol achieves competitive delivery probability, efficient resource utilization, and low overhead compared to the Epidemic and MaxProp protocols. Notably, it consistently outperformed the Epidemic protocol across all metrics and exhibited comparable delivery probability to MaxProp while maintaining significantly lower overhead, particularly in low-density scenarios. The results demonstrate the protocol’s ability to adapt to varying network conditions, effectively balance forwarding and resource management, and maintain robust performance in dynamic vehicular environments.

A portable EEG signal acquisition system and a limited-electrode channel classification network for SSVEP

Brain-computer interfaces (BCIs) have garnered significant research attention, yet their complexity has hindered widespread adoption in daily life. Most current electroencephalography (EEG) systems rely on wet electrodes and numerous electrodes to enhance signal quality, making them impractical for everyday use. Portable and wearable devices offer a promising solution, but the limited number of electrodes in specific regions can lead to missing channels and reduced BCI performance. To overcome these challenges and enable better integration of BCI systems with external devices, this study developed an EEG signal acquisition platform (Gaitech BCI) based on the Robot Operating System (ROS) using a 10-channel dry electrode EEG device. Additionally, a multi-scale channel attention selection network based on the Squeeze-and-Excitation (SE) module (SEMSCS) is proposed to improve the classification performance of portable BCI devices with limited channels. Steady-state visual evoked potential (SSVEP) data were collected using the developed BCI system to evaluate both the system and network performance. Offline data from ten subjects were analyzed using within-subject and cross-subject experiments, along with ablation studies. The results demonstrated that the SEMSCS model achieved better classification performance than the comparative reference model, even with a limited number of channels. Additionally, the implementation of online experiments offers a rational solution for controlling external devices via BCI.

A Load‐Balancing Enhancement to Schedule‐Aware Bundle Routing

ABSTRACTDelay‐ and disruption‐tolerant networking (DTN) enables communication in networks afflicted by long propagation delays and sporadic connectivity. DTN routing techniques such as schedule‐aware bundle routing (SABR) exist to route data bundles in deterministic networks, such as those found in deep‐space environments, where node contacts are predictable. This article begins with an overview of DTN architecture and SABR. SABR's method of final route selection (forwarding rules) is closely examined. The article then addresses a limitation of SABR whereby the algorithm may overlook parallel channels, leading to network congestion. To mitigate this, an enhancement is proposed. This enhancement aims to optimize data bundle distribution across candidate routes in networks with parallel channels, thus alleviating congestion and enhancing overall network performance. This is achieved with simple modifications to SABR's forwarding rules to avoid the concentration of data bundles on a minority of node contacts. The enhancement is demonstrated through simulations in a reference scenario implemented in DtnSim.

Evaluation of Variable Frequency Drive System in Pumping Stations of Irrigation Networks Using INACSEM Model

ABSTRACTLimited resources for the construction of new irrigation networks and low performance of most of them have led to more attention to evaluate and improve the performance of networks. The use of automated systems to improve performance and increase flexibility in irrigation networks is expanding. Regarding the variety of equipment, algorithms and applications of various automated systems in different parts of irrigation networks, it is necessary to develop performance evaluation models that include all aspects of automation. In this research, the irrigation networks automatic control system evaluation model (INACSEM) has been prepared using the classical evaluation method to assess automated control systems in irrigation networks. In this model, evaluation includes both technical and general approach. The technical point of view includes the types of equipment, software and automated control structures and their application in different stages of operation in different parts of the network and the general approach, which includes the general aspects of evaluation, mainly the managerial, social, economic and environmental point of views. The remote‐control system [by variable frequency drive (VFD)] installed in the pumping stations of Moghan plain was examined in order to show the application of the developed model. The results of performance evaluation of the automatic system show a performance of 82% with a validity of 74% in the pumping stations of Moghan plain. The results indicate the suitability of the model for evaluating automated systems in irrigation networks. Considering all aspects of automation, it also shows the problems of its components.

Deep Learning Method for Airfoil Flow Field Simulation Based on Unet++

ABSTRACTThis paper investigates the accuracy of U‐Net++ networks in predicting Reynolds‐Averaged Navier‐Stokes (RANS) solutions. The study employs the symbolic distance function (SDF) to represent geometry and flow conditions, utilizing parameterized airfoil data from the UIUC (University of Illinois at Urbana‐Champaign) airfoil datasets. The research assesses the performance of multiple trained neural networks in predicting pressure and velocity distributions. Specifically, the study examines the influence of varying network weights on solution accuracy. Through the optimization of the model, the research demonstrates that the mean relative error is below 1.72% for a range of previously unseen wing shapes, with a computational speedup factor of up to 1,000× in certain scenarios. The accuracy achieved by this model underscores the significant potential of deep learning‐based approaches as reliable tools for aerodynamic design and optimization.

Programmable packet-optical network security and monitoring using DPUs with embedded GPUs [Invited

Data processing units (DPUs) with embedded graphics processing units (GPUs) have the potential to revolutionize optical network functionalities at the edge. These advanced units can significantly enhance the performance and capabilities of optical networks by integrating powerful processing capabilities directly at the network edge, where data is generated and consumed. We explore the use cases for DPUs in optical data monitoring with local artificial intelligence (AI) processing and embedded security. This paradigm shift aims to enable more efficient data handling, reduced latency, and improved overall network performance by leveraging local AI processing capabilities embedded within DPUs. In this paper, we show how DPUs can analyze vast amounts of optical data in real-time, implementing advanced data analysis algorithms and security protocols directly on the DPUs to provide robust monitoring and protection for the optical networks. Results indicate that DPUs with embedded GPUs can significantly improve the detection and response times to network anomalies, performance issues, and security threats.

A comparative analysis using LEACH protocol to enhance energy efficiency in wireless sensor networks with harmony search algorithm

Wireless sensor networks (WSNs) are critical to many applications, but their use is frequently hampered by energy constraints, particularly in remote locations. WSN nodes are typically installed in remote and frequently inaccessible locations, making battery replacement challenging, thus emphasizing the importance of energy efficiency. This research presents an energy-conserving approach centred on the Harmony Search Algorithm (HSA), optimized by considering key parameters such as the number of sensor nodes, initial energy levels, transmission range, and data rate. The network performance is tested against the LEACH algorithm by evaluating parameters including energy consumption, network lifetime, and throughput. The comparative analysis is performed using graphs depicting the energy consumption of nodes by both approaches. With the Harmony Memory Size (HMS), Harmony Memory Considering Rate (HMCR), and Pitch Adjustment Rate (PAR) adjusted for best outcomes, the suggested method maximizes the performance of WSNs by effectively regulating the energy consumption of sensor nodes. Unlike LEACH, which involves clustering of nodes and periodic rotation of clusters, HSA is based on energy-efficient harmony and does not rely on clustering, thereby finding the most energy-efficient communication path within the network. Simulation results indicate that the HSA algorithm is 9.52% more efficient, with its energy usage being 9.2% more economical compared to the LEACH algorithm. This demonstrates HSA's potential for enhancing WSN performance in energy-constrained environments.

Coherent Optics for Passive Optical Networks: Flexible Access, Rapid Burst Detection, and Simplified Structure

With the development of the Internet of Things, cloud networking, and 4K/8K high-definition video, global internet traffic has seen a dramatic increase. This surge in traffic has placed higher demands on the performance of optical networks, featuring higher data rates, lower latency, and lower cost. The passive optical network (PON) is a representative scenario of optical access networks. Issues such as burst-mode detection in upstream PON scenarios, flexible rate allocation in downstream scenarios, and the simplification of hardware complexity at the optical network unit (ONU) side have attracted considerable attention. Compared to intensity modulation/direct detection (IM/DD), a recently proposed coherent PON incorporates a local oscillator laser at the receiver, enabling superior receiver sensitivity, spectrally efficient modulation, linear optical field recovery, and flexible channel selection. These features significantly enhance the flexibility and data rates of PON systems. This paper provides a comprehensive review of the development of coherent PONs, particularly in aspects of preamble design for burst-mode detection in upstream scenarios, the design of flexible rate PONs in downstream scenarios, and solutions for reducing hardware complexity at the ONU side.

Quantization tolerant network design and performance estimation of computation-in-memory for energy-efficient 3D object detection inference

Abstract This work discusses network architecture, network training method, and quantization method for achieving a low-energy and high-energy-efficient system, aiming at equipping robots and automobiles such as Autonomous Mobile Robot (AMR) with Computation-in-Memory (CiM) and performing 3D object detection with low energy consumption. Augmented Point Cloud VoxelNet (APCVN) is a network that improves inference accuracy by allowing slight increase in computational complexity. Multi-Stage Quantization Aware Training (MSQAT) and U-Quantization (UQ) are a learning method and a quantization strategy, respectively, that improve quantization tolerance of APCVN. Furthermore, quantitative calculations are conducted to estimate ADC energy consumption per inference, energy efficiency, memory array area, memory capacity, and latency per inference in assumed CiM system. Results show that by applying proposed methods, ADC energy consumption per inference is reduced by 8.7% and energy efficiency is improved by 1.6 times while maintaining high inference accuracy.

Pretrained Deep Neural Network Kin-SiM for Single-Molecule FRET Trace Idealization.

Single-molecule fluorescence resonance energy transfer (smFRET) has emerged as a pivotal technique for probing biomolecular dynamics over time at nanometer scales. Quantitative analyses of smFRET time traces remain challenging due to confounding factors such as low signal-to-noise ratios, photophysical effects such as bleaching and blinking, and the complexity of modeling the underlying biomolecular states and kinetics. The dynamic distance information shaping the smFRET trace powerfully uncovers even transient conformational changes in single biomolecules both at or far from equilibrium, relying on trace idealization to identify specific interconverting states. Conventional trace idealization methods based on hidden Markov models (HMMs) require substantial a priori knowledge of the system under study, manual intervention, and assumptions about the number of states and transition probabilities. Here, we present a deep learning framework using long short-term memory (LSTM) to automate the trace idealization, termed Kin-SiM. Our approach employs neural networks pretrained on simulated data to learn high-order correlations in the multidimensional FRET trajectories. Without user input of Markovian assumptions, the trained LSTM networks directly idealize the FRET traces to extract the number of underlying biomolecular states, their interstate dynamics, and associated kinetic parameters. On benchmark smFRET data sets, Kin-SiM achieves a performance similar to conventional HMM-based methods but with less hands-on time and lower risk of bias. We further systematically evaluate the key training factors that affect network performance to define the correct hyperparameter tuning for applying deep neural networks to smFRET data analyses.

A multi-level feature fusion artificial neural network for classification of acoustic emission signals.

In this paper, we introduce FUSION-ANN, a novel artificial neural network (ANN) designed for acoustic emission (AE) signal classification. FUSION-ANN comprises four distinct ANN branches, each housing anindependent multilayer perceptron. We extract denoised features of speech recognition such as linear predictive coding, Mel-frequency cepstral coefficient, and gammatone cepstral coefficient to represent AE signals. These features are concatenated to form a new feature called LMGC, which serves as input data for the four branches of FUSION-ANN. The network performs AE signal recognition and classification through forward propagation in each branch, utilizing multi-level feature fusion. We evaluate FUSION-ANN's performance on the ORION-AE benchmark dataset, which contains AE signals from various loading conditions simulating loosening phenomena in aeronautics, automotive, and civil engineering structures. Our results demonstrate an impressive average accuracy of 98% in AE signal classification. Additionally, FUSION-ANN boasts high training efficiency, robustness, and accuracy, making it suitable for reliable AE signal analysis. However, given the current limitations, we aim to conduct more comprehensive investigations in the future. Our plan includes further testing of the network's performance across various categories of AE signals to assess its generality. Additionally, we will select richer and more efficient feature sets to characterize these signals.

Design of multi-stage spectrum sensing via layer customized convolutional neural network for cooperative communication

ABSTRACT Cognitive radios make extensive use of cooperative spectrum sensing to make optimum use of the accessible spectrum. Multi-stage spectrum sensing in cooperative communication refers to a technique used in cognitive radio systems where multiple stages of sensing are employed to improve the accuracy spectrum sensing. Accordingly, this paper proposes a unique design of ISDP-AL-CNN-based multi-stage spectrum sensing for cooperative communication. The multi-stage spectrum sensing is performed under two stages: (i) Determining the presence of PU and (ii) Detect the availability of unused radio spectrum. In stage 1, the energy detection approach is performed for each received signal and then compare the energy detected signal with threshold to determine whether PU is present or not. In stage 2, the features such as SNR and PSD are extracted from the received signal and subjected to obtain feature vector and fed the formulated feature into Fusion centre to extract intense features. Further, the retrieved features are given to the ISDP-AL-CNN model to train the features. The ISDP-AL-CNN model adopts ISDP-based attention layer for better performance of the training network. Moreover, it uses SVD to set the data label in the ISDP-AL-CNN model and proficiently detects the spectrum availability in the CR-IoT network.

A Novel 5G Fronthaul Architecture Based on Quantum Security Protection

In the digital age, 5G technology has significantly enhanced global communication capabilities, providing strong support for various industries. However, 5G falls short of achieving comprehensive intelligence within the Internet of Things (IoT), propelling the development of 6G technology. 6G is expected to further enhance network performance by integrating advanced technologies such as AI and quantum communication, while expanding application scenarios to include communication in extreme environments for comprehensive global connectivity. This paper proposes an innovative fronthaul architecture that applies Quantum Key Distribution (QKD) technology to the 5G fronthaul, leveraging its unconditionally secure key generation and distribution mechanism. In this design, costlier Alice devices are placed on the AAU side, while less expensive Bob devices are positioned on the DU side, optimizing deployment to reduce engineering costs and lower deployment barriers. The feasibility of this architecture is demonstrated by calculating the secure key rate. Comparative analysis with existing research is performed to clarify future research directions. These innovations not only enhance 5G network security but also offer new solutions for the security requirements of 6G networks, such as data security, network resilience, and algorithm transparency. Our research offers strategic value for future network security, laying the groundwork for reliable networks.

GTD‐CER: Game‐Theoretic Dynamic Clustering for Enhanced Efficiency and Reliability in Wireless Sensor Networks

ABSTRACTEfficient cluster formation is a significant issue in wireless sensor networks (WSNs) for optimizing energy consumption, enhancing data reliability, and prolonging network lifetime. This paper proposes a dynamic cluster formation scheme modeled as a noncooperative game, where each sensor node acts as a player making strategic decisions on whether to become a cluster head (CH) or remain a member node (MN). In this role selection decision‐making phase, nodes evaluate their potential utility function, which incorporates energy consumption, data reliability, and data forwarding efficiency. Nodes dynamically switch roles based on their utility to achieve an optimal network configuration. Simulations conducted using the NS‐3 network simulator demonstrate that the proposed game theory‐based clustering scheme significantly improves network performance. We analyze key measures like energy consumption, packet delivery ratio (PDR), and data forwarding, and the results shows that the proposed scheme reduces energy consumption by 25%, improves PDR by 15%, and extends the network lifetime by 20%, compared to traditional methods. The dynamic role‐switching mechanism prolongs network lifetime by preventing the rapid depletion of node energy, ensuring stable and efficient network operations. It significantly enhances network performance, stability, and longevity.

Runoff Simulation and Waterlogging Analysis of Rainstorm Scenarios with Different Return Periods on Campus: A Case Study at China University of Geosciences

Urban flooding disasters are increasingly prevalent because of global climate change and urbanization. University campuses, as independent functional zones, exhibit complex rainfall–runoff dynamics. This study focuses on the China University of Geosciences, using data from two extremely heavy rainfall events and on-site waterlogging investigations in Wuhan in 2020 and 2021. A stormwater management model was employed to simulate campus catchment runoff and pipe network performance under rainstorm scenarios of various return periods, illustrating the spatial and temporal evolution of waterlogging on the campus. The simulation results indicate that the discharge at the main outlets aligned with rainfall patterns but exhibited a delayed response. During an overload period exceeding one hour, the ratios of overflow nodes and overloaded conduits reached 72.22% and 57.94%, respectively. Ponding was concentrated mainly in the southwest region of the campus, with the maximum ponding depth reaching 0.5 m. Future flood mitigation measures, such as enhancing permeable surfaces, upgrading pipeline infrastructure, and promoting rainwater reuse, could support the development of a “sponge campus” layout to alleviate flood pressure and enhance campus sustainability and resilience.

Cross-alignment of silver nanowires network for efficient nanowelding

The performance of silver nanowire (AgNW) network flexible transparent electrodes is limited by large contact resistance, making it necessary to perform nanowelding to improve conductivity of the network. However, not all nanowire junctions can be welded. Our work indicates that the welding kinetics between nanowires depend on the crossing angle, with higher surface diffusion velocity prone to welding and fracture at nanowire junctions of crossing angles close to 90 degrees. The impact of nanowire crossing angles on the welding process makes it difficult to achieve simultaneous welding of random AgNWs networks. To address this issue, we adopted an improved Meyer rod coating method to prepared a cross-aligned nanowire network based on a layer-by-layer assembly strategy. Compared to randomly distributed AgNWs networks (11.17 Ω sq-1, 85.2%), the cross-aligned AgNWs network achieved simultaneous welding of nanowire junctions during thermal annealing, further enhancing the optoelectronic performance (10.8 Ω sq-1, 90.3%) of the AgNWs network, resulting in a superior figure of merit value of 421.

Harnessing spatiotemporal transformation in magnetic domains for nonvolatile physical reservoir computing.

Combining physics with computational models is increasingly recognized for enhancing the performance and energy efficiency in neural networks. Physical reservoir computing uses material dynamics of physical substrates for temporal data processing. Despite the ease of training, building an efficient reservoir remains challenging. Here, we explore beyond the conventional delay-based reservoirs by exploiting the spatiotemporal transformation in all-electric spintronic devices. Our nonvolatile spintronic reservoir effectively transforms the history dependence of reservoir states to the path dependence of domains. We configure devices triggered by different pulse widths as neurons, creating a reservoir featured by strong nonlinearity and rich interconnections. Using a small reservoir of merely 14 physical nodes, we achieved a high recognition rate of 0.903 in written digit recognition and a low error rate of 0.076 in Mackey-Glass time series prediction on a proof-of-concept printed circuit board. This work presents a promising route of nonvolatile physical reservoir computing, which is adaptable to the larger memristor family and broader physical neural networks.

A Distributed Energy-Throughput Efficient Cross-Layer Framework Using Hybrid Optimization Algorithm

Magnetic induction (MI)-operated wireless sensor networks (WSNs), due to their similar performance in air, underwater, and underground mediums, are rapidly emerging networks that offer a wide range of applications, including mine prevention, power grid maintenance, underground pipeline monitoring, and upstream oil monitoring. MI-based wireless underground sensor networks (WUSNs), utilizing small antenna coils, offer a viable solution by providing consistent channel conditions. The cross-layer protocols address the specific challenges of WUSNs, leading to improved network performance and enhanced operational capabilities in real-world applications. This work proposes a distributed cross-layer solution, leveraging the hybrid marine predator naked mole rat algorithm (MPNMRA) for MI-operated WUSNs. The solution, called DECMN (distributed energy-throughput efficient cross-layer network using MPNMRA), is designed to optimize the MI communication channels, MI relay coils (MI waveguide), and MI waveguide with 3D coils to fulfill quality of service (QoS) parameters, while achieving energy savings and throughput gains. DECMN utilizes the interactions between various layers to develop cross-layer protocols based on MPNMRA. Simulation results demonstrate the effectiveness of DECMN, offering energy savings, increased throughput, and reliable transmissions within the performance limits.

Evaluating the effect of noise reduction strategies in CT perfusion imaging for predicting infarct core with deep learning.

This study evaluates the efficacy of deep learning models in identifying infarct tissue on computed tomography perfusion (CTP) scans from patients with acute ischemic stroke due to large vessel occlusion, specifically addressing the potential influence of varying noise reduction techniques implemented by different vendors. We analyzed CTP scans from 60 patients who underwent mechanical thrombectomy achieving a modified thrombolysis in cerebral infarction (mTICI) score of 2c or 3, ensuring minimal changes in the infarct core between the initial CTP and follow-up MR imaging. Noise reduction techniques, including principal component analysis (PCA), wavelet, non-local means (NLM), and a no denoising approach, were employed to create hemodynamic parameter maps. Infarct regions identified on follow-up diffusion-weighted imaging (DWI) within 48hours were co-registered with initial CTP scans and refined with ADC maps to serve as ground truth for training a data-augmented U-Net model. The performance of this convolutional neural network (CNN) was assessed using Dice coefficients across different denoising methods and infarct sizes, visualized through box plots for each parameter map. Our findings show no significant differences in model accuracy between PCA and other denoising methods, with minimal variation in Dice scores across techniques. This study confirms that CNNs are adaptable and capable of handling diverse processing schemas, indicating their potential to streamline diagnostic processes and effectively manage CTP input data quality variations.