In an industrial environment, sound is generally an important source of data that results from the movement and vibration of machines during operation. Each machine operation condition often generates a specific sound, for example, a familiar normal working sound, an unfamiliar sound, a breaking sound, or even the absence of sound, which may indicate a stop. This paper proposes a supervised convolutional neural network (CNN) model with early stopping and learning rate (LR) scheduling to achieve a more accurate and cost-effective trained deep learning (DL) model for machine sound classification. The proposed model is trained and evaluated using the MIMII dataset, which contains unbalanced real-world data and signal-to-noise ratios (SNRs). To ensure a fair (i.e., representative) data splitting strategy and robust coverage across all hierarchical structures of the MIMII dataset, this study adopts a hierarchical multi-stage stratified splitting protocol incorporating pre-split shuffling to prevent order bias. The synthetic minority oversampling technique (SMOTE) is used to balance the training set after feature extraction. Compared with state-of-the-art works implemented on the same dataset, the proposed model achieves optimal performance in terms of accuracy, area under the curve (AUC) , F1-score, precision, recall, matthews correlation coefficient (MCC), and Kappa without noise processing methods at an SNR of 6 dB, and maintains robust, competitive performance under severe noise conditions (0 dB and −6 dB).
To solve the limitations of traditional two-level inverters to medium-to-high-power photovoltaic (PV) systems, such as increased switching stress, electromagnetic interference, and total harmonic distortion (THD) as well as high cost and complicated structure, multilevel inverters are suggested to be more viable. In this paper, two typical five-level inverters potential to be used with PV cells, including a flying capacitor-based inverter and an active neutral point clamped (ANPC) inverter, are analyzed and compared. The control technique based on phase disposition modulation and redundant state selection is adopted, ensuring that the capacitor voltages in the active converter are balanced and controlled correctly. By simulating the system with MATLAB, the results show that both inverters are capable of reducing the THD to extremely low levels, contributing to output waves with improved waveform characteristics before and after filtering. Moreover, the waveform quality of the ANPC inverter is better with lower THD even without a filter applied to the output. This successfully demonstrates the feasibility of the implemented control strategy.
This paper proposes two types of passive auxiliary injection circuits (PAICs) that enable pulse tripling in three parallel-connected rectifiers, addressing the limitation of pulse multiplication to a factor of 2. The proposed design combined three rectifier units with two PAICs consisting of a delta/star transformer and nine auxiliary diodes. By further integrating the PAICs with conventional topologies such as 3-pulse star, 6-pulse star, 6-pulse bridge, 18-pulse star, and 18-pulse bridge rectifiers, we constructed 9-pulse star, 18-pulse star, 18-pulse bridge, 54-pulse star, and 54-pulse bridge configurations, respectively. The validation in MATLAB/Simulink demonstrated that these configurations achieved a threefold increase in both output-voltage pulses and input-current steps without the need for complex phase-shifting transformers. Moreover, the total harmonic distortion of the input current was significantly reduced, with values of 12.63%, 3.66%, 3.43%, 2.24%, and 1.90% for the respective designed rectifiers. To the best of our knowledge, this is the first demonstration of a passive pulse-tripling circuit for three parallel-connected rectifiers, offering a simple solution for high-current industrial applications.
In this work, an efficient AC-DC converter based on a bridgeless single-ended primary inductor converter (SEPIC) is proposed. Composed of only two anti-series switches, two diodes, two inductors, and one coupling capacitor, SEPIC has two different voltage outputs with inverted polarities, which makes it simpler and more attractive than the existing ones. More importantly, each output voltage can be individually adjusted by two different control loops, thus contributing to two different and independent duty cycles. With no need for a front-end diode bridge rectifier (DBR), its efficiency is enhanced. The simulation results on MATLAB/Simulink successfully validate that the converter is capable of supplying an output current in both directions. By connecting two independent loads to the output terminals, each with load characteristics of 100 V, 200 W, and 2 A which fit well with the application requirements of Fuji compact inverters, stable output voltages with a voltage ripple lower than 4.2% were obtained. The operating efficiency under rated load conditions was 95%, resulting in a 10% reduction in DBR usage. Moreover, the envelope of the input line current closely resembled the standard sinewave with a total harmonics distortion (THD) of 2.00%, while the AC load current had a THD of 6.89%, both within standard limits.
The intelligent transportation systems require secure, low-latency, and reliable communication architectures to enable the real-time vehicular application. This paper proposes an edge-intelligent semantic aggregation (EISA) framework for 6G unmanned aerial vehicle (UAV)-assisted Internet of vehicles (IoV) networks that integrates task-driven semantic communication, deep reinforcement learning (DRL)-based edge intelligence, and blockchain-based semantic validation across 6G terahertz (THz) links. UAVs in the proposed architecture serve as adaptive edge nodes that receive semantically vital information about the vehicle at any given stage, optimize aggregation and transmission parameters dynamically, and guarantee data integrity through a structured, lightweight consortium blockchain that signs semantically detailed representations rather than raw packets. Simulation results from a hybrid NS-3, MATLAB, and Python environment indicate that the proposed framework can achieve up to 45% reduction in end-to-end latency, an approximately 70% increase in throughput, and semantic efficiency with blockchain verification delays of less than 20 ms (more than 98%). These findings support the effectiveness of the proposed co-design for achieving context-aware, energy-efficient, and reliable communication under heavy-traffic conditions. The proposed framework provides a flexible and scalable foundation for next-generation 6G-enabled automotive networks, with subsequent growth toward federated learning-based collaborative intelligence, digital-twin-assisted traffic modeling, and quantum-safe blockchain mechanisms to enhance scalability, intelligence, and long-term security.
In this paper, a comprehensive evaluation on the silicon/silicon carbide (Si/SiC) hybrid switch is performed through by experimental tests in terms of both electrical performance and robustness under extreme stresses. Based on the optional turn-on and turn-off delay times under the efficiency control mode obtained from the double-pulse test (DPT), both nondestructive and destructive single-pulse avalanche tests are conducted on the Si/SiC hybrid switch as well as on the two discrete device branches inside the hybrid switch. In addition, the avalanche voltage, critical avalanche energy, and peak avalanche current, which intrinsically characterize the unclamped-inductive-switching (UIS) avalanche characteristics, are carefully examined. In this way, the physical factors dominating the UIS characteristics of the hybrid switch, thus limiting its single-pulse avalanche withstand capability, are specifically and comprehensively identified; the underlying physical mechanisms are analyzed and revealed in depth, and how the gate control sequence affects the UIS characteristics of the hybrid switch is extensively investigated. We additionally carry out short-circuit (SC) tests under the fault-under-load (FUL) condition and perform a parallel in-depth analysis to experimentally determine which branch dominates the SC withstand capability of the hybrid switch. Our experimental study indicates that, for both SC robustness and single-pulse avalanche capability, the limiting factor is a single device branch among the two parallel discrete devices, and the UIS behavior is sensitive to the variation of the gate turn-off delay time Toff_delay. The study conducted in this paper not only provides deep academic insights into the electrical performance and reliability of the Si/SiC hybrid switch, but also offers fundamental theoretical principles and technical evidence to support more efficient and long-term reliable applications of the hybrid switch in the industrial fields.
This study presents a comprehensive theoretical investigation of the electronic and structural properties of a series of fractal molecular architectures derived from benzene and progressively extended toward circumcoronene-like graphene analogues. Molecular geometries were constructed using GaussView 6, while all quantum-chemical calculations were carried out within the Gaussian 09 package using density functional theory (DFT) at the B3LYP/6-31G level, ensuring a reliable balance between computational accuracy and efficiency. The gradual expansion of the hexagonal framework resulted in a systematic reduction in the energy gap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO), indicating enhanced electronic delocalization and improved charge-transport characteristics in higher-order fractal structures. Electron density contour maps revealed an increased π-electron symmetry in the central regions of the larger systems, whereas smaller units exhibited an edge-localized electron density, consistent with the development of extended π–π conjugation. In addition, the density of states (DOS) spectra demonstrated a pronounced broadening of both occupied and virtual states with an increasing structural size, confirming the strong correlation between the fractal growth and electronic transport behavior. Furthermore, analysis of the HOMO and LUMO distributions showed an orbital broadening and enhanced spatial symmetry in advanced fractal geometries, accounting for the observed reduction in the energy gap. These results indicate that coronene-based fractal structures exhibit significant potential for applications in conductive nanomaterials and molecular electronics, as their electronic and structural properties can be finely tuned through controlled fractal branching, enabling tailored performance in next-generation nanoscale devices.
Owing to superior breakdown voltage and excellent robustness, the beta-gallium oxide (β-Ga2O3) power device has emerged as a pivotal research frontier in power electronics. Although advanced packaging strategies, including nano-silver paste sintering, alumina direct bond copper (DBC) substrates, and flip-chip structures, have been adopted to mitigate the intrinsic low thermal conductivity of β-Ga2O3. However, a further reduction in the thermal resistance while maintaining high reliability remains a challenge. This study introduces a novel packaging methodology that synergistically integrates nano-silver films with aluminum nitride active metal brazing (AMB-AlN) substrates, achieving an ultra-low junction-to-case thermal resistance. By comprehensive reliability assessments on β-Ga2O3 Schottky barrier diodes (SBDs) and hetero-junction diodes (HJDs), the results demonstrate that the SBDs and HJDs exhibit surge current densities of 0.876 kA/cm2 and 0.778 kA/cm2, respectively, which represents a significant advancement in device performance benchmarks. These advancements provide critical insights into packaging design for high-reliability ultrawide bandgap semiconductor systems.
With the growing deployment of unmanned aerial vehicles (UAVs) swarms in national defense, military operations, and emergency response, secure and reliable intra-swarm identity authentication has become critical for ensuring coordinated action and mission reliability. To address the drawbacks of public key infrastructure (PKI) based authentication in UAV swarms, namely, complex certificate management, strong dependence on centralized authorities, and authentication latency. We proposed a certificateless identity authentication scheme for UAV swarms built on blockchain sharding. The scheme leverages sharding to execute authentication in parallel across multiple shards, significantly improving efficiency. Each UAV locally generates its public/private key pair and then adopts a registration-based encryption (RBE) mechanism: A registration algorithm binds the device identity to its key on the blockchain, ensuring public verifiability and immutability of identity mapping. On this basis, an authentication algorithm runs in which the initiator produces an authentication signature using a common reference string (CRS), on-chain public-key registration information, and its local private key, and the verifier rapidly validates the authentication message using the on-chain registration data and the identity of the initiator. The experimental results demonstrate that the proposed scheme achieves low-latency, high-throughput identity authentication in large-scale UAV swarm environments, providing a solid technical foundation and broad application prospects for trustworthy UAV swarm identity authentication.
Zero-day attacks present a critical cybersecurity challenge for Internet of things (IoT) infrastructures, where the inability of signature-based intrusion detection systems (IDSs) to recognize novel threat behaviors compromises both system reliability and operational continuity. Existing hybrid IDS solutions often struggle to balance accurate classification of known attacks with reliable anomaly detection, particularly under the computational constraints of IoT environments. To address this gap, we introduce ZeroDefense, an adaptive fusion-based IDS designed for simultaneous detection of known intrusions and emerging zero-day threats. The framework employs a four-layer architecture consisting of i) feature standardization and class balancing, ii) anomaly detection using isolation forest, autoencoder, and local outlier factor, iii) fine-grained attack classification via random forest, extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), and attentive interpretable tabular learning (TabNet), and iv) a confidence-aware fusion engine that adaptively selects the most reliable decision path. Suspicious or previously unseen traffic is isolated early through fused anomaly scoring, while benign and known-malicious flows are processed through supervised classification for precise attack labeling. With an anomaly cascaded decision pipeline, a dynamic confidence-driven fusion mechanism, and a deployment-conscious design, ZeroDefense enables real-time inference on IoT edge gateways. Evaluation on the CICIoT2023 benchmark demonstrates 99.94% overall accuracy and 95.64% macro-average F1-score for known attacks, while 5.76% of traffic is successfully flagged as potential zero-day activity, with inference latency maintained below 100 ms/flow. These results indicate that ZeroDefense offers a scalable, resilient, and practically deployable defence capability for modern IoT infrastructures.
Due to the complex structural hierarchy, with deeply nested associative relationships between entities such as equipment, specifications, and business processes, intelligent power grid engineering is challenging. Meanwhile, limited by the fragmented data and loss of contextual information, the generated reports are prone to the problems such as content redundancy and omission of critical information, failing to meet the demands of efficient decision-making and accurate management in modern power systems. To address these issues, this paper proposes a knowledge graph (KG)-enhanced framework to automatically generate electric power engineering reports. In the KG construction phase, a feature-fused entity recognition model named BERT-BiLSTM-CRF is adopted to improve the accuracy of entity recognition in scenarios involving power engineering professional terminology, thereby solving the problem of ambiguous entity boundaries in traditional models; then a BERT-attention relation extraction model is proposed to enhance the completeness of extracting complex hierarchical and implicit relationships in power grid data. In the report generation phase, an improved Transformer architecture is adopted to accurately transform structured knowledge into natural language reports that comply with engineering specifications, addressing the issue of semantic inconsistency caused by the loss of structural information in existing models. By validating with real-world projects, the results show that the proposed framework significantly outperforms existing baseline models in entity recognition, confirming its superiority and applicability in practical engineering.
Side-channel analysis (SCA) has emerged as a research hotspot in the field of cryptanalysis. Among various approaches, unsupervised deep learning-based methods demonstrate powerful information extraction capabilities without requiring labeled data. However, existing unsupervised methods, particularly those represented by differential deep learning analysis (DDLA) and its improved variants, while overcoming the dependency on labeled data inherent in template analysis, still suffer from high time complexity and training costs when handling key byte difference comparisons. To address this issue, this paper introduces invariant information clustering (IIC) into SCA for the first time, and thus proposes a novel unsupervised learning-based SCA method, named IIC-SCA. By leveraging mutual information maximization techniques for automatic feature extraction of power leakage data, our approach achieves key recovery through a single training session, eliminating the prohibitive computational overhead of traditional methods that require separate training for all possible key bytes. Experimental results on the ASCAD dataset demonstrate successful key extraction using only 50000 training traces and 2000 attack traces. Furthermore, compared with DDLA, the proposed method reduces training time by approximately 93.40% and memory consumption by about 6.15%, significantly decreasing the temporal and resource costs of unsupervised SCA. This breakthrough provides new insights for developing low-cost, high-efficiency cryptographic attack methodologies.
Single-phase non-isolated microinverters used in photovoltaic (PV) systems commonly encounter two persistent challenges: High-frequency leakage current and fluctuating power delivery. This paper presents a novel single-phase, non-isolated multi-input microinverter topology with a common-ground structure that effectively eliminates ground leakage current without requiring additional active components. The proposed microinverter architecture integrates a dual-boost configuration and uses only four active switches. This is especially advantageous in terms of the component count, which is beneficial to enhance reliability, reduce cost, and simplify the overall system design. With one, two, or four PV inputs, it can operate without interruption under unbalanced voltage or partial shading and even if some inputs drop to zero. A tailored modulation scheme minimizes conduction losses while maintaining a stable direct-current (DC)-link voltage, and a decoupling capacitor efficiently absorbs the single-phase pulsating power, thus overcoming one major limitation in existing microinverter designs. By validating with a 1-kW GaN-based prototype, both the simulated and experimental results demonstrate its high efficiency, robustness, and practical suitability for cost-effective PV applications, with a peak efficiency value of 94.8%.
In the era of rapidly expanding wireless technologies, the push for larger spectrum efficiency and better signal integrity has intensified the need for high-efficient and low-noise amplifiers (LNAs). A two-stage LNA based on the GaAs/InGaAs pseudomorphic high electron mobility transistor (pHEMT) with a relatively large gate length of 2 μm is designed for high-performance 2.4-GHz wireless communication. The I-V characteristic and two-port high-frequency S-parameter of the transistor are measured by on-wafer probing techniques. The results indicate that a discrete transistor with a gate size of can provide a maximum trans-conductance of 16 mS, corresponding to a maximum current-gain cut-off frequency and maximum oscillation frequency of 7 GHz and 8 GHz, respectively, at the 1-V drain-source voltage. With the impedance matching networks based transmission line technique, an extended integrated layout structure is designed and simulated by using the momentum simulation tool embedded in Advanced Design System (ADS), targeting to alleviate the trade-off between noise figure (NF) and gain of the circuit. The findings show that the transistor based on the GaAs/InGaAs technology is capable of delivering high performance with power consumption low to 16 mW, where the maximum simulated gain and minimum NF of 21.5 dB and 2.4 dB are achieved, respectively. In terms of linearity, the proposed LNA provides terrific output 1-dB compression and output third-order intercept point values of –3 dBm and 10 dBm, respectively. The bandwidth of 0.12 GHz and figure-of-merit (FOM) of 12 are obtained, which are comparable to that of the previously reported LNAs based on pHEMT. Such a device may benefit to accelerate the development of more robust and power-efficient front-end modules in modern wireless systems, especially for advancing performance-driven applications.
Neural network-based methods for intrapulse modulation recognition in radar signals have demonstrated significant improvements in classification accuracy. However, these approaches often rely on complex network structures, resulting in high computational resource requirements that limit their practical deployment in real-world settings. To address this issue, this paper proposes a Bottleneck Residual Network with Efficient Soft-Thresholding (BRN-EST) network, which integrates multiple lightweight design strategies and noise-reduction modules to maintain high recognition accuracy while significantly reducing computational complexity. Experimental results on the classical low-probability-of-intercept radar signal dataset demonstrate that BRN-EST achieves comparable accuracy to state-of-the-art methods while reducing computational complexity by approximately 50%.
This article presents a compact crab-shaped reconfigurable antenna (CSRA) designed for 5G Sub-6 GHz wireless applications. The antenna achieves enhanced gain in a miniaturized form factor by incorporating a hexagonal split-ring structure controlled via two Radio Frequency (RF) PIN diodes (BAR64-02V). While the antenna is primarily designed to operate at 3.50 GHz for Sub-6 GHz 5G applications, RF switching enables the CSRA to cover a broader frequency spectrum, including the S-band, X-band, and portions of the Ku-band. The proposed antenna offers several advantages: it is low-cost (fabricated on an FR-4 substrate), compact (achieving 64.07% size reduction compared to conventional designs), and features both frequency and gain reconfigurability through digitally controlled PIN diode switching. The reflection coefficients of the antenna, both without diodes and across all four switching states, were experimentally validated in the laboratory using a Keysight FieldFox microwave analyzer (N9916A, 14 GHz). The simulated radiation patterns and gain characteristics closely matched the measured values, demonstrating an excellent agreement. This study bridges the gap between traditional and next-generation antenna designs by offering a compact, cost-effective, and high-performance solution for multiband, reconfigurable wireless communication systems. The integration of double-split-ring resonators and dynamic reconfigurability makes the proposed antenna a strong candidate for various applications, including S-band and X-band systems, as well as the emerging lower 6G band (7.125–8.40 GHz).
The rapid and increasing growth in the volume and number of cyber threats from malware is not a real danger; the real threat lies in the obfuscation of these cyberattacks, as they constantly change their behavior, making detection more difficult. Numerous researchers and developers have devoted considerable attention to this topic; however, the research field has not yet been fully saturated with high-quality studies that address these problems. For this reason, this paper presents a novel multi-objective Markov-enhanced adaptive whale optimization (MOMEAWO) cybersecurity model to improve the classification of binary and multi-class malware threats through the proposed MOMEAWO approach. The proposed MOMEAWO cybersecurity model aims to provide an innovative solution for analyzing, detecting, and classifying the behavior of obfuscated malware within their respective families. The proposed model includes three classification types: binary classification and multi-class classification (e.g., four families and 16 malware families). To evaluate the performance of this model, we used a recently published dataset called the Canadian Institute for Cybersecurity Malware Memory Analysis (CIC-MalMem-2022) that contains balanced data. The results show near-perfect accuracy in binary classification and high accuracy in multi-class classification compared with related work using the same dataset.
The convergence of Internet of things (IoT) and 5G holds immense potential for transforming industries by enabling real-time, massive-scale connectivity and automation. However, the growing number of devices connected to the IoT systems demands a communication network capable of handling vast amounts of data with minimal delay. These generated enormous complex, high-dimensional, high-volume, and high-speed data also brings challenges on its storage, transmission, processing, and energy cost, due to the limited computing capabilities, battery capacity, memory, and energy utilization of current IoT networks. In this paper, a seamless architecture by combining mobile and cloud computing is proposed. It can agilely bargain with 5G-IoT devices, sensor nodes, and mobile computing in a distributed manner, enabling minimized energy cost, high interoperability, and high scalability as well as overcoming the memory constraints. An artificial intelligence (AI)-powered green and energy-efficient architecture is then proposed for 5G-IoT systems and sustainable smart cities. The experimental results reveal that the proposed approach dramatically reduces the transmitted data volume and power consumption and yields superior results regarding interoperability, compression ratio, and energy saving. This is especially critical in enabling the deployment of 5G and even 6G wireless systems for smart cities.
To fulfill the training requirements for the daily operations of multirotor Unmanned Aerial Vehicles (UAV) clusters, a UAV cluster collaborative task integrated simulation platform (UAV-TISP), was developed. The platform integrates a suite of hardware and software to simulate a range of collaborative UAV cluster operation scenarios. It features modules for collaborative task planning, UAV cluster simulations, and tactical monitoring. The platform significantly reduces training costs by eliminating physical drone dependencies while offering a flexible environment for testing swarm algorithms. UAV-TISP supports both individual UAV and swarm operations, incorporating high-fidelity flight dynamics, real-time communication via user datagram protocol (UDP), and collision avoidance strategies. Utilizing the OSGEarth engine, it enables dynamic 3D environment visualization and scenario customization. Three key task scenarios—route flight, formation reconstruction, and formation transformation—were tested to validate the platform’s efficacy. Results demonstrated robust formation maintenance, adaptive collision avoidance, and seamless task execution. Comparative analysis with Gazebo Sim revealed lower trajectory deviations in UAV-TISP, highlighting its superior accuracy in simulating real-world flight dynamics. Future work will focus on enhancing scalability for diverse UAV models, optimizing swarm networking under communication constraints, and expanding mission scenarios. UAV-TISP serves as a versatile tool for both operational training and advanced algorithm development in UAV cluster applications.
With the advancement of electronic countermeasures,airborne synthetic aperture radar(SAR)systems are facing increasing challenges in maintaining effective performance in hostile environments.In particular,high-power interference can severely degrade SAR imaging and signal processing,often rendering target detection impossible.This highlights the urgent need for robust anti-interference solutions in both the signal processing and image processing domains.While current methods address interference across various domains,techniques such as waveform modification and spatial filtering typically increase the system costs and complexity.To overcome these limitations,we propose a novel approach that leverages the multi-domain characteristics of interference to efficiently suppress narrowband interference and repeater modulation interference.Specifically,narrowband interference is mitigated using notch filtering,a signal processing technique that effectively filters out unwanted frequencies,while repeater modulation interference is addressed through strong signal amplitude normalization,which enhances both the signal and image processing quality.These methods were validated through tests on real SAR data,demonstrating significant improvements in the imaging performance and system robustness.Our approach offers valuable insights for advancing anti-interference technologies in SAR systems and provides a cost-effective solution to enhance their resilience in complex electronic warfare environments.
