Search Results (931)

The jacking renovation construction of aging bridges faces significant safety risks due to the complexity and uncertainty of their structures. Addressing the limitations of traditional risk assessment methods in handling dynamic changes and data scarcity, this study proposes a safety risk assessment approach based on dynamic Bayesian networks (DBN) and fuzzy set theory (FST). By using DBN to model the temporal evolution of risks, combined with the Leaky Noisy-OR Gate extension model and FST to quantify expert knowledge, this method overcomes the constraints of insufficient data. Taking an elevated bridge jacking renovation project in Qingdao, China, as a case study, a risk indicator system was established, incorporating factors such as personnel, equipment, and the environment. The results show that risks are higher in the early stages of construction and stabilize later on, with poor foundation conditions, instability of the substructure, and improper operations identified as key risk sources requiring focused control. Through forward reasoning, the study predicts risk trends, while backward reasoning identifies sensitive factors, providing a scientific basis for construction safety management. Full article

The operational complexity of wastewater treatment systems mainly stems from the diversity of influent characteristics and the nonlinear nature of the treatment process. Together, these factors make the control of effluent quality in wastewater treatment plants (WWTPs) difficult to manage effectively. To address this challenge, constructing accurate effluent quality models for WWTPs can not only mitigate these complexities, but also provide critical decision support for operational management. In this research, we introduce a deep learning method that fuses multi-source data. This method utilises various indicators to comprehensively analyse and predict the quality of effluent water: water quantity data, process data, energy consumption data, and water quality data. To assess the efficacy of this method, a case study was carried out at an industrial effluent treatment plant (IETP) in Anhui Province, China. Deep learning algorithms including long short-term memory (LSTM) and gated recurrent unit (GRU) were found to have a favourable prediction performance by comparing with traditional machine learning algorithms (random forest, RF) and multi-layer perceptron (MLP). The results show that the R² of LSTM and GRU is 1.36%~31.82% higher than that of MLP and 9.10%~47.75% higher than that of traditional machine learning algorithms. Finally, the RReliefF approach was used to identify the key parameters affecting the water quality behaviour of IETP effluent, and it was found that, by optimising the multi-source feature structure, not only the monitoring and management strategies can be optimised, but also the modelling efficiency of the model can be further improved. Full article

As a critical storage technology, the material selection and structural design of flash memory devices are pivotal to their storage density and operational characteristics. Although van der Waals materials can potentially take over the scaling roadmap of silicon-based technologies, the scaling mechanisms and optimization principles at low-dimensional scales remain to be systematically unveiled. In this study, we experimentally demonstrated that the floating-gate length can significantly affect the memory window characteristics of memory devices. Experiments involving various floating-gate and tunneling-layer configurations, combined with TCAD simulations, were conducted to reveal the electrostatic coupling behaviors between floating gate and source/drain electrodes during shaping of the charge storage capabilities. Fundamental performance characteristics of the designed memory devices, including a large memory ratio (82.25%), good retention (>50,000 s, 8 states), and considerable endurance characteristics (>2000 cycles), further validate the role of floating-gate topological structures in manipulating low-dimensional memory devices, offering valuable insights to drive the development of next-generation memory technologies. Full article

This paper proposes a 1200V 4H-SiC MOSFET incorporating a High-K dielectric-integrated fused source-gate (HKSG) structure, engineered to concurrently enhance the third-quadrant operation and high-frequency figure of merit (HF-FOM). The High-K dielectric enhances the electric field effect, reducing the threshold voltage of the source-gate. As a result, the reverse conduction voltage drops from 2.79 V (body diode) to 1.53 V, and the bipolar degradation is eliminated. Moreover, by incorporating a shielding area within the merged source-gate architecture, the gate-to-drain capacitance C_gd of the HKSG-MOS is reduced. The simulation results show that the HF-FOM C_gd × R_on,sp and Q_gd × R_on,sp of the HKSG-MOS are decreased by 48.1% and 58.9%, respectively, compared with that of conventional SiC MOSFET. The improved performances make the proposed SiC MOSFEET have great potential in high-frequency power applications. Full article

This paper presents a smart electric vehicle (EV) charging management system that integrates Reinforcement Learning intelligence on a Field-Programmable Gate Array (FPGA) platform. The system is based on the Q-learning algorithm, where the RL agent perceives environmental conditions, captured through hardware sensors such as current, voltage, and priority indicators, and makes optimal charging decisions to address grid stress and prioritize charging needs. The FPGA implementation leverages hardware design strategies to ensure efficient operation and real-time response within a limited amount of required energy, allowing for its implementation in embedded applications and possibly enabling the use of an energy harvesting power source, like a small solar panel. The proposed design effectively manages multiple EV chargers by dynamically allocating current and prioritizing charging tasks to maintain service quality. Through intelligent decision making, informed by continuous sensor feedback, the system adapts to fluctuating grid conditions and optimizes energy distribution. Key findings highlight the system’s ability to maintain stable operation under varying demand conditions, improving power efficiency, safety, and service reliability. Moreover, the design is scalable, enabling seamless expansion for larger installations by following consistent architectural guidelines. This FPGA-based solution combines RL intelligence, sensor-based environmental perception, and robust hardware design, offering a practical framework for an efficient EV charging infrastructure in modern smart grid environments. Full article

As the primary public source of satellite trajectory data, the Two-Line Element (TLE) dataset offers fundamental orbital parameters for space missions. However, for satellites with poor data quality, traditional neural network models often underperform, hindering accurate orbit predictions and meeting demands in satellite operation and space mission planning. To address this, a federated-learning-based trajectory prediction enhancement strategy is proposed. Satellites with low training efficiency and similar orbits are grouped for collaborative learning. Each satellite uses a Convolutional Neural Network (CNN) model to extract features from historical prediction error data. The server optimizes the global model through the Federated Averaging algorithm, learning more orbital patterns and enhancing accuracy. Experimental results confirm the method’s effectiveness, with a marked increase in prediction accuracy compared to traditional methods, validating federated learning’s advantage. Moreover, the combination of federated learning with basic neural network models like the Multi-Layer Perceptron (MLP), Long Short-Term Memory (LSTM), Recurrent Neural Network (RNN), and Gated Recurrent Unit (GRU) is explored. The results indicate that integrating federated learning can greatly enhance satellite prediction, opening new possibilities for future orbital prediction and space technology development. Full article

This paper presents two cryogenic low-noise amplifiers (LNAs) based on the WIN’s 0.18 $\mathrm{\mu }$m gate length gallium arsenide (GaAs) pseudomorphic high electron mobility transistor (pHEMT) process designed for radio telescope receivers. Discrete transistors with gate peripheries spanning 50–600 $\mathrm{\mu }$m were DC-characterized at 290 K and 15 K, respectively. The LNAs underwent on-chip noise characterization under 15 K using a Y-factor measurement setup, which integrated a calibrated noise source and a noise figure analyzer. This approach directly quantified the noise temperature—critical metrics for radio telescope receiver front-ends. The top-performing LNA variant identified through on-chip characterization was packaged and evaluated in a cryogenic test-bed. This LNA, spanning a bandwidth of 0.3–15 GHz, demonstrated a gain of 26 dB and a minimum noise temperature of 6 K when operated at an ambient temperature of 15 K. In contrast, a second LNA architecture, tested solely on-chip, demonstrated a gain of 30 dB and a minimum noise temperature of 15 K across the 0.3–7 GHz range. Full article

Digital pulse-width modulation (DPWM)-based controls are characterized by a non-negligible phase delay due to analog-to-digital (ADC) conversion, sampling time, carrier shape, and algorithm computation time. These delays degrade the performance in closed-loop systems, where the bandwidth must be reduced to avoid instability issues due to the reduced closed-loop phase margin. To mitigate these delays, approaches such as utilizing low-latency ADCs or increasing the sampling frequency have been employed. However, these methods are often costly and do not address the fundamental delay issues inherent to DPWMs. In this paper, a novel zero phase-delay DPWM architecture is proposed. This enhanced architecture seamlessly integrates pulse width and frequency modulation to create a programmable derivative action, capable of effectively recovering the DPWM delay. The proposed architecture employs a reliable and straightforward organization, suitable for implementation in commercial field programmable gate array (FPGA). Furthermore, this architecture inherently generates a trigger signal that can be used in numerous power electronic applications to capture the average value in piecewise linear inductor currents. The validity of the proposed architecture is substantiated through simulations and experimental tests. The final implementation is shared in an open-source resource. Full article

The Qinghai–Tibet Plateau (QTP), a critical hydrological regulator for Asia through its extensive glacier systems, high-altitude lakes, and intricate network of rivers, exhibits amplified sensitivity to climate-driven alterations in precipitation regimes and ice mass balance. While the Gravity Recovery and Climate Experiment (GRACE) and its Follow-On (GRACE-FO) missions have revolutionized monitoring of terrestrial water storage anomalies (TWSAs) across this hydrologically sensitive region, spatial resolution limitations (3°, equivalent to ~300 km) constrain process-scale analysis, compounded by mission temporal discontinuity (data gaps). In this study, we present a novel downscaling framework integrating temporal gap compensation and spatial refinement to a 0.25° resolution through Gated Recurrent Unit (GRU) neural networks, an architecture optimized for univariate time series modeling. Through the assimilation of multi-source hydrological parameters (glacier mass flux, cryosphere–precipitation interactions, and land surface processes), the GRU-based result resolves nonlinear storage dynamics while bridging inter-mission observational gaps. Grid-level implementation preserves mass conservation principles across heterogeneous topographies, successfully reconstructing seasonal-to-interannual TWSA variability and also its long-term trends. Comparative validation against GRACE mascon solutions and process-based hydrological models demonstrates enhanced capacity in resolving sub-basin heterogeneity. This GRU-derived high-resolution TWSA is especially valuable for dissecting local variability in areas such as the Brahmaputra Basin, where complex water cycling can affect downstream water security. Our study provides transferable methodologies for mountainous hydrogeodesy analysis under evolving climate regimes. Future enhancements through physics-informed deep learning and next-generation climatology–hydrology–gravimetry synergy (e.g., observations and models) could further constrain uncertainties in extreme elevation zones, advancing the predictive understanding of Asia’s water tower sustainability. Full article

Under wideband interference conditions, traditional neural networks often suffer from low accuracy in single-frequency direction-of-arrival (DOA) estimation and face challenges in detecting single-frequency sound sources. To address this limitation, we propose a novel model called Parallel Net. The architecture adopts a frequency-parallel design: it first employs a recurrent neural network, the generalized feedback gated recurrent unit (GFGRU), to independently extract features from each frequency component, and then it fuses these features through an attention mechanism. This design significantly enhances the network’s capability in estimating the DOA of single-frequency signals. The simulation results demonstrate that when the signal-to-noise ratio (SNR) exceeds −10 dB, Parallel Net achieves a mean absolute error (MAE) below 2°, outperforming traditional frequency-coherent neural networks and the MUSIC algorithm, and reduces the error to half that of classical beamforming (CBF). Further validation on the SWellEx-96 experiment confirms the model’s effectiveness in detecting single-frequency sources under wideband interference. Parallel Net exhibits superior sidelobe suppression and fewer spurious peaks compared to CBF, achieves higher accuracy than MUSIC, and produces smoother and more continuous DOA trajectories than conventional neural network models. Full article

Qudits allow the encoding and manipulation of additional quantum information compared to that stored to a two-level qubit system. Although manipulations of qudit states are generally more complex and can introduce extra sources of noise, qudits can still be used in a number of applications when this error can be kept sufficiently low. One such application is the case of the Bucket Brigade Algorithm for realizing a Quantum RAM (QRAM), which inherently uses qutrits for encoding the state of address switches. In this paper, we study a methodology for qutrit manipulation that leverages efficient encoding techniques and pulse calibration methods for the case of transmon systems. The methodology employs an encoding scheme that allows the execution of controlled operations, using the subspace spanned by the two lowest levels of the transmon; we show how this scheme can be used for generating one- and two-qutrit gates by leveraging the Qiskit and Boulder Opal frameworks to compute the parameters of pulses that implement the quantum gates that are used by the BBA. For this type of gate, simulations show that the pulses perform the required operations with a low infidelity when errors introduced by the qutrit Hamiltonian dynamics are considered. Full article

Recently, multiple new technologies have emerged for automating the development of neural network (NN) accelerators for both field-programmable gate arrays (FPGAs) and application-specific integrated circuits (ASICs). This paper explores methodologies for translating NN algorithms into chip layouts, with a focus on end-to-end automation, cost-effectiveness, and open-source software. We present a robust framework for developing NN-to-silicon solutions and demonstrate a seamless plug-and-play automation flow using TensorFlow, Vivado HLS, HLS4ML, and Openlane2. SkyWater Technologies’ 130 nm PDK (Sky130) is employed to successfully generate layouts for two small NN examples under 1000 parameters, incorporating dense, activation, and 2D convolution layers. The results affirm that current open-source tools effectively automate low-complexity neural network architectures and deliver faster performance through FPGA structures. However, this improved performance comes at the cost of increased die area compared to bare-metal designs. While this showcases significant progress in accessible NN automation, achieving manufacturing-ready layouts for more complex NN architectures remains a challenge due to current tool limitations and heightened computational demands, which points to exciting opportunities for future advancements. Full article

Small non-coding molecules known as microRNAs (miRNAs) play a critical role in disease diagnosis, treatment, and prognosis evaluation. Traditional wet-lab methods for validating miRNA–disease associations are often time-consuming and inefficient. With the advancement of high-throughput sequencing technologies, deep learning methods have become effective tools for uncovering potential patterns in miRNA–disease associations and revealing novel biological insights. Most of the existing approaches focus primarily on individual molecular behavior, overlooking interactions at the multi-molecular level. Conventional graph neural network (GNN) models struggle to generalize to heterogeneous graphs, and as network depth increases, node representations become indistinguishable due to over-smoothing, resulting in reduced predictive performance. GONNMDA first integrates similarity features from multiple data sources and applies noise reduction to obtain a reconstructed, comprehensive similarity representation. It then constructs heterogeneous graphs and applies a root–tree hierarchical alignment, along with an ordered gating message-passing mechanism, effectively addressing the challenges of heterogeneity and over-smoothing. Finally, a multilayer perceptron is employed to produce the final association predictions. To evaluate the effectiveness of GONNMDA, we conducted extensive experiments where the model achieved an AUC of 95.49% and an AUPR of 95.32%. The results demonstrate that GONNMDA outperforms several recent state-of-the-art methods. In addition, case studies and survival analyses on three common human cancers—breast cancer, rectal cancer, and lung cancer—further validate the effectiveness and reliability of GONNMDA in predicting miRNA–disease associations. Full article

The delimitation of an ecotone is an important reference for ecosystem conservation; however, the assessment of a marine ecotone from an ecological point of view represents a knowledge gap. The Yellow River Estuary (YRE) serves as both spawning and feeding grounds for numerous economically important organisms. Delineating the boundary of YRE and assessing the boundary change have great importance in maintaining its ecosystem health. This study attempts to apply a Moving Split Window (MSW) to determine marine boundary in YRE. Level 2 remote sensing satellite data spanning from 2012 to 2020 sourced from the Geostationary Ocean Color Imager (GOCI) were utilized. Chlorophyll-a, Chromophoric Dissolved Organic Matter (CDOM), and Total Suspended Solids (TSS) were employed as variables, with Squared Euclidean Distance (SED) serving as the determinant for identifying the marine ecological ecotone within the Yellow Estuary and its adjacent waters. Results indicate the following: (1) SED values exhibit distinct peaks and valleys, facilitating the accurate identification of marine ecotones via MSW. (2) Evident ecotones are observable in both the gate and coastal regions. (3) The influence range of TSS on the gate spans between 10 km and 14 km. In synthesis, the ensuing conclusions are drawn: MSW proves to be a reliable method for quantitatively determining ecotones in marine environments. Furthermore, MSW introduces a novel approach to the delineation of marine ecotones. Full article

This study aims to assess the environmental impact of using wood-based biomass as a high-efficiency fuel alternative to fossil fuels for heat production. To achieve this, the life cycle of biomass transformation, utilization, and disposal was analyzed using the life cycle assessment (LCA) methodology with SimaPro 9.5.0.2 PhD software. The system boundaries included extraction, processing, transportation, combustion, and waste management, following a cradle-to-gate approach. A comparative analysis was conducted between natural gas, the most widely used conventional heating fuel, and two biomass-based fuels: wood pellets and wood chips. The results indicate that biomass utilization reduces greenhouse gas emissions (−19%) and fossil resource depletion (−16%) while providing environmental benefits across all assessed impact categories analyzed, except for land use (+96%). Biomass is also to be preferred for forest waste management, ease of supply, and energy independence. However, critical life cycle phases, such as raw material processing and transportation, were found to contribute significantly to human health and ecosystem well-being. To mitigate these effects, optimizing combustion efficiency, improving supply chain logistics, and promoting sustainable forestry practices are recommended. These findings highlight the potential of biomass as a viable renewable energy source and provide insights into strategies for minimizing its environmental footprint. Full article