A knowledge graph-guided 3D visual question answering method was proposed to capture the implicit common-sense semantic relationships between objects in the scene. By introducing external structured knowledge, the model was effectively enhanced in both semantic understanding and reasoning ability. Specifically, key semantic entities were extracted from the question text, and a knowledge graph-guided feature enhancement module was designed to obtain knowledge features using these key semantic entities. The knowledge features were fused with visual features extracted both from the question representation and from the 3D object detection network for answer prediction. Experimental results on the ScanQA dataset showed that the proposed method outperforms existing baseline models on metrics such as EM@1 and BLEU-4.

A railway foreign object detection algorithm for complex weather conditions based on an improved RT-DETR, named FRP-DETR, was proposed. A feature complementary mapping module (FCM) and Pzconv units were introduced, which built complementary paths between shallow spatial details and deep semantic information to compensate for the limitations of single-scale feature representation, with the two components working collaboratively to enhance the perception capabilities of small and edge targets. A railway perception modulation fusion module (RMFM) was introduced, which was based on adaptive channel attention and spatial modulation mechanisms to enhance the model’s response to key semantic information in railway scenes. The original downsampling module was replaced with pinwheel-shaped convolution (PSConv), which enhanced edge texture extraction through multi-directional asymmetric padding and separable convolution. A railway foreign object detection dataset containing four weather conditions was constructed based on the SaMam style transfer method, by transferring sunny railway foreign object images to rainy, foggy, and snowy scenes. Experimental results showed that compared to the original RT-DETR-R18 model, this method achieved improvements of 1.65 percentage points and 3.4 percentage points in mAP@0.5 and mAP@0.5:0.95, respectively, with a 63.5% reduction in parameter count and an inference speed of 88 frames per second. The results verified that FRP-DETR achieved high accuracy, lightweight design, and real-time performance in railway foreign object detection under complex weather conditions.

Significant progress in multimodal large language models (MLLMs) has driven advances in visual question answering, visual understanding, and reasoning tasks, and their potential for deployment on resource-constrained edge devices is increasingly recognized. However, large model sizes and the substantial costs of deployment and inference remain major barriers to practical adoption. Optimizing MLLMs for edge devices has become a critical research direction in this field. A comprehensive survey of recent advances in optimizing MLLMs for edge deployment was presented, along with the associated challenges and development trends. The research evolution of MLLMs on edge devices was reviewed, with particular emphasis on model architecture optimization and inference scheduling strategies. In model architecture optimization, techniques including visual information compression, sparse attention, and mixture-of-experts models were specifically analyzed. System-level optimizations involving computation scheduling, hardware adaptation, compilation optimization, and cloud-edge collaboration were investigated to enhance inference efficiency and energy efficiency. Furthermore, the key challenges of these models in practical applications were discussed, and a variety of task scenarios ranging from assistive to collaborative and autonomous types were covered, categorized by the perspective of autonomy levels. Finally, current limitations were summarized and future research directions regarding standardized deployment, efficient computing and storage, and multi-modal fusion optimization were outlined.

To address the issue of inconsistent thermal stress between the upper and lower devices of the same bridge arm in converters under conditions of uneven aging or layout deviations, a novel thermal balancing method for upper and lower bridge arms based on dead-time shifting adjustment was proposed. By monitoring the on-state voltage of insulated gate bipolar transistor (IGBT) online, the real-time junction temperature of IGBTs was deduced based on the mapping relationship between the on-state voltage and junction temperature of IGBTs. Then, the temperature difference between the upper and lower devices of the same bridge arm was determined. When the thermal stress difference between the upper and lower bridge arms exceeded the threshold, the effective conduction time of the upper and lower bridge arms was adjusted by fine-tuning the position of the dead time, thereby actively regulating the power loss distribution of the upper and lower bridge arms. This achieved an increase in the loss of the bridge arm with lower temperature and a decrease in the loss of the bridge arm with higher temperature without affecting the total system loss, thus improving the thermal stress consistency of power devices in the converter and enhancing system reliability. Theoretical analysis and experimental results showed that the proposed method did not affect system efficiency and had no significant negative impact on output performance. Under typical operating conditions, the average temperature difference between the upper and lower devices was reduced by approximately 14% after adjustment, effectively improving the thermal stress consistency of the upper and lower bridge arms.

Current industrial image anomaly detection methods generally face challenges such as high dependency on anomalous samples, insufficient realism of synthetic samples, and limited perception capability for complex defects. To An industrial image anomaly detection method based on diffusion synthesis and feature mining was proposed to address these issues. Accordingly, a category-sensitive selective diffusion anomaly synthesis module was designed to generate pseudo-anomaly samples with adjustable intensity and category adaptability through controllable perturbation and category-sensitive loss, which could effectively alleviate data scarcity. Meanwhile, a multi-stage feature mining framework was constructed, including contrast-driven feature selection, multi-dimensional perception attention reconstruction, and residual refinement selection modules, enabling dynamic screening of anomaly-sensitive features and enhancement of structural details. Experimental results demonstrated that the proposed method achieved outstanding performance on the MVTec AD and MPDD datasets, with image-level AUROC scores of 99.7% and 98.4%, and pixel-level AUROC scores of 99.0% and 98.7%, respectively, validating its effectiveness and robustness.

A detection model named dynamic kernel perception YOLO (DKP-YOLO) was constructed to address the problems of contextual information loss, multi-scale distress coexistence, and insufficient detail restoration in pavement distress detection from a UAV perspective. A lightweight crack context module was proposed to mitigate the variability of crack morphology and the loss of contextual information in complex scenes. This module enhanced feature discriminability through heterogeneous parallel depthwise separable convolutions and a dynamic feature fusion mechanism. In order to tackle the challenges of complex backgrounds and multi-scale distress, a pavement distress perception module was designed, which improved the multi-scale distress perception capability using parallel multi-scale convolutions and a dual-attention mechanism. A large-kernel perception feature fusion module was introduced to overcome the limitations of traditional convolutions in modeling long-range dependencies and responding to small targets. This module captured global contextual relationships by integrating extra-large receptive field convolution with channel-spatial attention. Finally, a lightweight dynamic upsampling operator was incorporated into the neck network to reduce detail loss and semantic ambiguity during upsampling. Experiments on the UAV-PDD2023 dataset showed that the proposed model significantly improved detection accuracy while effectively reducing model complexity. Compared with the baseline model, DKP-YOLO achieved a better balance between being lightweight and maintaining high detection performance, making it more suitable for UAV-based pavement distress detection.

An encirclement decision-making method based on adaptive curriculum learning and multi-agent proximal policy optimization (MAPPO) algorithm was proposed to address the problem of encirclement and game confrontation for unmanned surface vehicle (USV) swarms in complex marine environments. For maritime encirclement and defense tasks, a combat simulation environment including dynamic targets and multiple reefs was constructed, and the judgement criteria for successful encirclement and task termination were defined. To improve the decision-making efficiency in complex adversarial environments, a normalized state space representing the motion states of both friendly and hostile parties, multi-scale reward functions, and a continuous action space were designed. An adaptive curriculum scheduler was introduced into the centralized training and decentralized execution framework to dynamically adjust the environmental complexity and the noise intensity of action exploration. The results of multiple sets of adversarial simulation experiments indicate that, compared to the baseline methods (i.e., no curriculum learning and traditional curriculum learning methods), the proposed method effectively improves the encirclement success rate, shortens the average task completion time and the average encirclement path length, reduces the number of collisions, and exhibits good generalization ability and adversarial adaptability.

Existing scheduling models for Internet data center (IDC) commonly have two key shortcomings: insufficient consideration of the voltage fluctuations in distribution networks and the impact of IDC storage data constraints on the spatiotemporal transfer of the workloads. An optimized operation method for flexible distribution networks was proposed considering the flexible IDC scheduling. Uncertainty scenarios for renewable generation and load demand were generated using Monte Carlo simulation and the K-means clustering algorithm. Differentiated scheduling strategy was established to distinguish between sensitive and tolerant loads. An IDC power consumption model along with a data storage constraint model were constructed. A robust optimization model was established with the objective of minimizing both distribution network operating costs and IDC electricity purchasing costs. The dynamic regulation capability of soft open point (SOP) and voltage fluctuation constraints were also incorporated into the model. Simulation results demonstrated that the proposed method significantly enhanced voltage quality, markedly increased the renewable energy accommodation rate, and reduced the electricity procurement cost for IDC by collaboratively optimizing the spatiotemporal load shifting of IDC workload along with the power regulation and reactive power support capabilities of SOP. It also validated the impact of storage capacity constraints on the efficiency of spatiotemporal workload shifting.

Autonomous driving simulation platforms play a vital role in the development, testing and validation of autonomous driving systems. A systematic review of the classification and key technical pathways of mainstream simulation platforms was presented, covering aspects such as environment modeling, sensor simulation, vehicle dynamics modeling, perception algorithm evaluation, V2X communication and cloud-based simulation. Core challenges and research progress related to synthetic data generation, cost-effective algorithm training, cross-domain generalization and platform scalability were analyzed, emphasizing simulation technologies based on computer vision and artificial intelligence. Future development trends were discussed. Simulation platforms are evolving toward higher realism, interactivity and closed-loop validation with the advancement of emerging technologies such as generative AI, neural rendering and multimodal learning, gradually forming a comprehensive pipeline that integrates data generation, algorithm training and performance evaluation. Simulation platforms will continue to play an essential role in enhancing the generalization capability of autonomous driving system, accelerating product deployment, and establishing standard testing and validation framework in the future.

Combustion experiments were conducted on spherical magnesium particles with a diameter of 3 mm in a methane flame environment. The effects of the volume fraction of effective oxidizing agent (13%?26%) and flame ambient temperature (1058?1220 K) on the combustion process of magnesium particles and the particle residue products were analyzed. The combustion of magnesium particles went through the stages of ignition, flame expansion, flame contraction, product aggregation and extinction. The outer layer of crystalline magnesium combustion residues exhibited a regular, loose, and white morphology. The inner layer of crystals was disorderly distributed and densely black. The composition of crystals was MgO. The combustion time was synergistically affected by the volume fraction of the effective oxidizing agent and the ambient temperature, primarily through diffusion and reaction kinetic mechanisms. The combustion time decreased with the increase of the volume fraction of effective oxidizing agent and the decrease of environmental temperature. A calculation formula for combustion time was obtained. The combustion time of 3 mm magnesium particles was sensitive to the volume fraction of effective oxidizing agent at low volume fraction. The effect of ambient temperature diminished with the increasing volume fraction of effective oxidizing agent.

A heterogeneous smoke and flame visual detection method integrating visible and infrared images was proposed, in order to address the problems of low detection efficiency, poor accuracy, and consequent delays in fire disposal and misallocation of resources in the early stage of tunnel fires caused by lighting interference and structural environment. For infrared images, an OTSU threshold segmentation optimized by a grey wolf optimizer fused with differential evolution was applied, and the maximum entropy algorithm was combined to improve flame region segmentation accuracy. Considering the tunnel scene characteristics, YOLO11 was adapted, and the backbone network was optimized using a Swin-Transformer modified with a neural attention mechanism to enhance weak feature extraction. During feature fusion, a lightweight BiFPN was employed to adapt to multi-scale smoke and flame features, and an adaptive WIoU loss function was introduced to improve localization accuracy under fuzzy boundaries. At the decision layer, weighted fusion and conflict resolution were used to achieve dual-modal collaborative decision-making. Experimental results demonstrated that the proposed method achieved at least a 23.55% reduction in average localization error compared with single-modal detection. Compared with nine fusion algorithms such as evidence theory, the recall, F1, and mAP values of the proposed method reached 92.52%, 90.83%, and 93.47%, respectively, all exceeding those of the compared algorithms, while maintaining the smallest localization error fluctuation. The inference speed of the model increased by an average of 14.61%. The research results not only contribute to tunnel fire prevention and control but also effectively improve overall tunnel safety management and emergency response capability.

A collaborative control approach integrating multi-agent proximal policy optimization (MAPPO) and maximum pressure control (MPC) was developed to address traffic control challenges at mixed traffic intersections involving connected autonomous vehicles (CAVs), human-driven vehicles (HDVs), and pedestrians. A hierarchical state space and an action space were designed through the formulation of a decentralized partially observable Markov decision process (Dec-POMDP). Reward functions considering the balance among safety, traffic efficiency, phase switching frequency, and regulatory compliance were introduced. The proposed model was validated under low, medium, and high traffic flow conditions using a centralized training with decentralized execution framework on the SUMO simulation platform. Experimental results demonstrated that the proposed MAPPO-MPC method significantly improved the throughput (by up to 44.3%) while reducing the queue length (by up to 49.3%), average vehicle delay (by up to 43.2%), and pedestrian waiting time (by up to 43.53%). Moreover, the model exhibited more substantial performance advantages when the CAV penetration rate exceeded 50%, outperforming the traditional Webster-based methods and the baseline models.

In response to the challenges of high difficulty in modeling and the low control efficiency in multivariable nonlinear complex systems, fuzzy control has been identified as a promising research approach. Significant progress has been made over the years. Advancements in fuzzy control theory for these systems have been reviewed. Firstly, relevant knowledge of fuzzy theory and the basic components of fuzzy controllers were presented. Secondly, various fuzzy models and control methods for multivariable nonlinear complex systems were discussed, including fuzzy control systems based on semantic rules, model-based fuzzy control systems, and others. Special emphasis was placed on model-based fuzzy control systems, such as the T-S fuzzy model, fuzzy cell-to-cell mapping model, fuzzy neural network model, and fuzzy models based on the variation relationship between process input and output data. Detailed summaries and analyses of both model structure identification and parameter identification were provided. Finally, challenges related to the universality, interpretability, stability, and applicability of these systems were explored, along with potential directions for future research and development.

A spatiotemporal joint lane-change intention prediction model based on Transformer and graph attention neural network (GATv2) was proposed in order to accurately predict lane-changing intention of surrounding vehicles and improve the safety and reliability of autonomous vehicles. Symmetric exponential average filtering was applied to smooth the raw NGSIM data. Then velocity and acceleration were recalculated by using the difference method, and outliers were removed. The data of the target vehicle and its surrounding vehicles were unified by establishing an eight-neighborhood for each vehicle, providing a foundation for model training and validation. Transformer multi-head attention was used to mine temporal dependency, while GATv2 was employed to quantify the interaction relationship between vehicles at each time step. Then a Transformer-GATv2 lane-changing prediction model based on spatiotemporal feature and multi-vehicle interaction was constructed. The experimental results showed that the proposed model significantly outperformed other baseline models in accuracy, recall, F₁-score, and other metrics. A prediction accuracy of 98.04% was achieved, which demonstrated superior lane-changing intention prediction performance and strong early prediction capability. The model was deployed on a six-degree-of-freedom driving simulator integrated with the CARLA co-simulation platform for model-in-the-loop testing, verifying the usability of the model in real-world scenario.

Discrete element method (DEM) and Tavares breakage model were employed to investigate the rock fragmentation mechanism and the statistical characteristics of fragments under impact loading. Drop hammer impact crushing process of limestone specimens with different aspect ratios (1.0—2.0) and sizes (diameter 30—50 mm) was systematically simulated. Image J image processing technology was used to quantitatively analyze the fragment count, particle size distribution, and shape indices. The results showed that increasing the aspect ratio and the size significantly raised the crushing energy threshold, which led to a delay in the peak normal force, a reduction in its amplitude, and a slowdown of the overall fragmentation process. The equivalent particle size of fragments showed poor alignment with the sieve analysis curve, which indicated that the fragments commonly exhibited irregular geometric configurations such as flaky and rod-like shapes. Fragments from specimens with lower aspect ratios (1.0) and smaller sizes (diameter 30 mm) had lower roundness, more pronounced angularity, and greater surface roughness. A quantitative relationship model among geometric parameters–crushing response–fragment morphology can provide theoretical basis and regulatory pathways for the design of impact-resistant structures, the optimization of blasting parameters, and the control of particle morphology in rockfill grading.

The Electric_Data inspection dataset was constructed, and a PDD-SPCIE algorithm based on context information enhancement and multi-target semantic perception was proposed within the RT-DETR framework, in order to improve the detection accuracy of multi-category power defects in complex substation inspection scenarios. The LDDM-ResNet module was designed to extract high- and low-frequency detailed features through multi-kernel parallel convolution. The DA-Encoder module was constructed to enhance the representation of multi-scale and irregularly shaped targets by incorporating deformable attention and dynamic sampling. The LGCA module was introduced to fuse global and local features using atrous convolution and spatial depth transformation convolution. The original framework was improved from the aspects of feature representation, context modeling, and feature fusion, thus enhancing the unified detection ability for multiple defect categories. Experimental results showed that the proposed model achieved a precision of 0.819, a recall of 0.681, and an mAP₅₀ of 0.704 on the self-built dataset, and a precision of 0.719, a recall of 0.642, and an mAP₅₀ of 0.712 on the VOC dataset.

A multi-stage surrogate model named LSTM-LFM-SRU was developed to address the high computational cost and limited real-time capability of traditional hydrodynamic models. The urban flood process was divided into two stages: one-dimensional drainage through pipe networks and two-dimensional surface inundation. Long short-term memory (LSTM) networks and a combination of a low-fidelity physical model (LFM) with a super-resolution U-shaped convolutional neural network (SRU) were adopted to replace the corresponding stages. A flood prediction framework was constructed by combining data-driven methods with physical information. The multi-stage surrogate model was applied to a case study in the Xiasha area of Hangzhou. A 20-hour intense rainfall-induced inundation event was simulated within 5.31, with its computational efficiency reaching 12.7 times that of the high-precision hydrodynamic models. The one-dimensional LSTM-based surrogate achieved an average Nash-Sutcliffe Efficiency value of 0.94, while the two-dimensional LFM-SRU model reduced the Root Mean Square Error by 17.2% compared with the LFM-U-net model.

A small object detection algorithm FMCM-YOLO based on feature enhancement and fusion attention mechanism was proposed, aiming at the challenges of limited feature extraction, foreground-background confusion, and severe missed and false detections in small object detection in optical remote sensing images. Firstly, a four-head detection model was designed and a small target detection layer was added to detect numerous small objects in optical remote sensing images. Secondly, a feature enhancement module was proposed in the backbone network, which improved feature extraction capability by designing a multi-branch convolutional structure and introducing dilated convolution of different sizes. Thirdly, channel and spatial attention mechanisms were incorporated into the neck network, and a residual structure was introduced to focus on small objects, facilitating the distinction between targets and backgrounds. Finally, MPDIoU was adopted as the model’s loss function to accelerate convergence and enhance detection performance for small objects. Experimental results demonstrated that the mAP50 of the proposed algorithm on the two public datasets, USOD and AI-TOD, reached 89.9% and 60.6% respectively, which were 2.8 and 5.9 percentage points higher than those of the baseline algorithm YOLOv5m. Especially, the mean average precision for extremely tiny, tiny, and small objects increased by 2.1, 6.5, and 5.1 percentage points, respectively. These results proved that the FMCM-YOLO algorithm effectively improved the detection performance of small targets in optical remote sensing images.

Existing knowledge graph update approaches often suffer from high computational cost, long update latency, and difficulties in maintaining consistency, which limit their applicability in large-scale dynamic environments. To address these issues, an incremental update method based on core entities and local subgraphs was proposed. A weighted centrality metric was constructed by integrating entity degree centrality with business access frequency, and the top-K entities with the highest scores were selected as core entities. In this way, the update process was focused on the entities that were most influential to downstream applications. Local subgraphs associated with the identified core entities were extracted through k-hop neighborhood expansion and edge-weight filtering, so that the update scope was restricted to semantically relevant structures and redundant computation was reduced. During model optimization, a local objective function was established for fine-tuning pre-trained knowledge graph embeddings. A neighborhood stability regularization term was introduced to suppress embedding drift caused by incremental updates. In addition, Schema constraints and multi-source confidence fusion were incorporated to ensure the consistency and reliability of the updated graph. Experiments were conducted on large-scale knowledge graph datasets. The results showed that, compared with full graph reconstruction and conventional fine-tuning methods, the proposed method reduced update latency by approximately 60%, improved triple insertion accuracy by approximately 8 percentage points, and decreased the consistency conflict rate by approximately 12 percentage points. These results indicate that the proposed method can effectively improve update efficiency while maintaining the structural and semantic quality of the knowledge graph.

Capturing characteristic modulation frequencies of propulsor noise can be challenging, as the actual acoustic signals in the scenario of passive acoustic detection exhibit extremely low signal-to-noise ratios. To tackle this challenge, an enhanced demodulation method, the Traversal Weighted Envelope Spectrum (TWES), was proposed, tailored for extracting the modulation frequencies of blade-type marine propulsor noise. First, the cyclostationary model for blade-type marine propulsor noise was established, and its cyclostationary signature was derived in detail. Then, the indicator Frequency-domain Correlated Kurtosis (FCK) was proposed, and a weighting function relying on prior knowledge of shaft frequency was constructed. Ultimately, the TWES was constructed for situations with prior knowledge of shaft frequency. Simulation results showed that the proposed TWES showed better demodulation performance than other benchmark methods even in the situation of SNR=?16 dB. The demodulation detection capability of TWES under low SNR levels was validated by experimental data from the pump jet and the propeller. The comparison with the classical demodulation methods and state-of-the-art demodulation methods further verified the superiority of the proposed TWES.