Search Results (1,465)

Small-object detection in unmanned aerial vehicle (UAV) imagery is hindered by limited target pixels, dense spatial distributions, background interference, and scale-sensitive bounding-box regression. This study develops a lightweight YOLO11n configuration in which established SCSA recalibration, DySample reconstruction, decoupled prediction, and SimOTA assignment act at successive stages of the detection pipeline. Its principal methodological contribution is a gradient-adaptive WIoU–NWD objective that uses previously observed regression-gradient fluctuations to balance overlap-oriented and distribution-based localization without altering the inference graph. On the official VisDrone2019-DET test-dev server, the 2.48 M-parameter model achieves 45.9% mAP@0.5, 27.0% mAP@0.5:0.95, and 20.2% AP${}_{\mathrm{small}}$, improving the YOLO11n baseline by 2.3, 1.4, and 2.4 percentage points, respectively; it also improves mAP@0.5/mAP@0.5:0.95 by 2.2/1.3 points on the vehicle-focused UAVDT benchmark and reaches 40.2 FPS on a Jetson Orin NX in 15 W mode using TensorRT FP16. The two benchmarks mainly represent urban, traffic, and low-altitude surveillance imagery; consequently, the cross-dataset result supports transfer within these conditions rather than universal generalization to all UAV applications. These results support a compact single-pass accuracy–efficiency trade-off for resource-constrained UAV perception, while the modest margin over fixed loss weighting indicates that the adaptive mechanism should be interpreted as an incremental, primarily small-object localization improvement rather than a complete solution to regression instability. Full article

Metro stations, while essential for urban transportation, pose unique evacuation challenges due to confined layouts and high densities; existing models often struggle to accurately capture individual pedestrian behaviors and the dynamic spread of fires. This study introduces a human–fire–agent interaction model designed to enhance the understanding and simulation of critical interactions among pedestrians, fire dynamics, and the underground environment of a metro station. The model integrates social force modeling and fluid dynamics to accurately represent pedestrian behavior and fire spread, for a more complete analysis of evacuation scenarios. Using Lumu Station in Suzhou as a case study, this study develops a detailed simulation framework implemented in an integrated PyroSim-Python-AnyLogic platform to model the evacuation process. The framework is employed to evaluate the effectiveness of turnstile reversal strategies—an approach that involves temporarily reversing the direction of turnstiles to facilitate faster evacuation during emergencies. Beyond mitigation, this study extends to the preparedness phase by functioning as a high-fidelity digital twin. It enables immersive “Serious Game” training and provides a quantitative tool for railway managers, decision-makers, and engineers to optimize operating procedures and performance-based station designs. Full article

Dam displacement forecasting is important for assessing whether long-term structural responses remain consistent with established operational behavior. In multi-point monitoring systems, however, irregular survey-line layouts and unequal numbers of monitoring points make it difficult to organize long-term records while preserving their engineering meaning. This study develops a field-oriented forecasting framework by reconstructing daily observations into a structured displacement-field object defined by survey-line order, monitoring-point alignment, and three displacement components. A valid-position-aware protocol is introduced to distinguish actual monitoring locations from structural padding, ensuring that model training and evaluation remain restricted to the same physical monitoring definition. Using long-term operational records from the Tianshengqiao First Dam, four representative models, namely SimVP, SimVPv2, PatchTST, and TimesNet, are evaluated under the same chronological split, causal forward-fill-only preprocessing, input window, prediction horizon, and evaluation boundary. All four models achieve strong predictive performance, with R² values above 0.97 in the X direction and above 0.99 in the Y and Z directions. No single trained model or forecasting route exhibits a consistent advantage across all displacement components and evaluation metrics. Under the present single-dam, case-specific setting, the relative ranking varies with displacement direction and forecasting horizon and should not be interpreted as evidence of general direction-specific suitability for any particular architecture. At the route level, the field-based route retains a slight advantage in Y-direction forecasting and overall MAE, whereas the sequence-based route remains competitive for Z-direction displacement and longer-horizon X-direction prediction. The proposed framework provides a practical and physically consistent digital representation for organizing irregular monitoring records, comparing forecasting routes, and supporting deployment-oriented model selection and subsequent model adaptation. Full article

Early and accurate plant disease detection is essential for reducing crop losses and supporting sustainable agricultural management. Although deep learning-based approaches have achieved strong performance in plant disease analysis, many existing models require substantial computational resources, which limits their use in resource-constrained agricultural environments. This study proposes ShuffleNetV2-hSimKD, a lightweight integration framework for plant disease detection. It adopts ShuffleNetV2 as the backbone and incorporates the parameter-free SimAM attention mechanism to enhance disease-related feature representation without introducing additional learnable parameters. In addition, the standard ReLU activation function is replaced with h-swish to improve nonlinear feature extraction and preserve informative feature responses. A hybrid knowledge distillation strategy is further employed to transfer both output-level and feature-level knowledge from a high-capacity teacher model to the lightweight student network during training. Unlike previous studies that apply these techniques in isolation, ShuffleNetV2-hSimKD synergistically integrates parameter-free SimAM, h-swish optimization, and hybrid KD to overcome the representation limitations of lightweight backbones in subtle disease symptom detection. The proposed framework was evaluated on a balanced subset of the PlantVillage dataset, in which leaf images were categorized as healthy or diseased. ShuffleNetV2-hSimKD achieved an accuracy of 90.41% with only 1.4M parameters and 151M FLOPs. Compared with representative lightweight Convolutional Neural Networks (CNNs), the proposed model achieved improved accuracy and recall while maintaining low computational complexity. These results demonstrate that ShuffleNetV2-hSimKD provides an effective balance between detection performance and computational efficiency, highlighting its potential as a lightweight candidate for plant disease detection in resource-constrained agricultural scenarios. Full article

This paper presents an efficient framework for surrogate modeling and rapid optimization of a dual-band circular patch antenna with a C-shaped slot (DB-CPAC) using multilayer perceptron (MLP) neural networks. Although highly accurate, traditional full-wave electromagnetic simulations are computationally expensive for geometric optimization due to complex slot-induced surface current perturbations. To address this limitation, a hybrid optimization framework based on Latin Hypercube Sampling (LHS) is proposed, combining the developed MLP model with a Method-of-Moments (MoM) simulator. The surrogate model uses an advanced modular architecture consisting of an ensemble of MLP neural networks for regressing center frequencies and classification MLP modules with a softmax output layer to estimate the probabilities of achieving bandwidth and gain targets. All networks are trained using the Levenberg–Marquardt algorithm with early stopping on data generated by a dedicated DB-CPAC_MoM_Sim software package. The proposed LHS-based optimizer employs the surrogate model for rapid global search and targeted local optimization before final MoM verification. Results show that this hybrid approach achieves an order-of-magnitude acceleration of the optimization process compared to conventional MoM methods while maintaining high accuracy. Full article

Large Language Model (LLM) agents in long-horizon planning often exhibit Semantic-Execution Drift (SED), where executed actions progressively deviate from original language constraints. To formalize this phenomenon, we model SED as a stochastic drift process, expressed by the recurrence D(t+1) = alpha D(t) + epsilon(t) + beta P(t), and show that policies with an alpha below one induce semantic contraction. We propose ReflectiChain, a framework integrating Retrospective Reflection, a Latent World Model, and Double-Loop Policy Adaptation to preserve semantic consistency during long trajectories. To evaluate SED, we introduce Sema-Sim, a multi-agent supply chain benchmark containing 10 policy constraints, six adversarial perturbations, and 30-step planning horizons. We further propose the Semantic Fidelity Index (SFI) for measuring instruction adherence. Experiments on DeepSeek-V3.2 across seven reasoning strategies show that ReflectiChain achieves the highest SFI (88.7) and stable semantic contraction (alpha = 0.823, below one). The results are consistently validated on Qwen3.5-122B and Qwen2.5-72B. Ablation studies demonstrate that Retrospective Reflection contributes most to performance gains. Additional analyses on scalability, failure modes, and cost efficiency further verify the robustness and practicality of the proposed framework. All code and evaluation resources are publicly released. Full article

Modern power grids require safer and more reliable field operations, yet conventional robots often face limitations in unstructured environments because of rigid pre-programming and weak perception–action coupling. This review examines Embodied Intelligence (EI) as an emerging direction for enhancing power-system field operations. We first evaluate the environmental adaptability of morphological carriers, including quadrupeds, humanoids, and unmanned aerial vehicles, and then define the perception–cognition–execution closed-loop architecture used in this review. Three application domains are then examined. Intelligent inspection focuses on active perception and potential open-vocabulary object detection. Live-line maintenance emphasizes Sim-to-Real methods and shared autonomy, while disaster-response applications involve heterogeneous air–ground robotic coordination. The review also discusses the potential for EI to reduce human exposure to hazardous tasks and influence labor structures, while a regional text-based proxy illustrates differences in policy attention to digital infrastructure. Finally, we analyze major constraints, including hardware endurance under extreme climates, edge-computing latency, foundation-model uncertainty and hallucination, cybersecurity, and safety certification. Overall, EI should not be interpreted as a mature replacement for current utility practice; it is a developing technological direction whose safe deployment will require field validation, standardized evaluation, cybersecurity assurance, and continued human supervisory authority. Full article

With advances in robotic manipulation in recent years, tactile sensing has become increasingly important in scenarios where visual information is unreliable or insufficient. However, the development of learning-based tactile algorithms and policies is limited by the cost and time required to collect large-scale physical datasets. While robotic simulation offers an alternative for synthetic data generation, current simulation platforms, such as NVIDIA Isaac Sim, lack integrated capacitive tactile sensors. This paper presents a finite element method (FEM)-based digital twin of a capacitive tactile sensor and an extension for NVIDIA Isaac Sim that enables the real-time generation of synthetic tactile data directly within the simulation environment. The proposed system extracts nodal deformations from the Isaac Sim PhysX engine and uses a convolutional neural network (CNN) to predict synthetic tactile maps that replicate the response of the physical sensor. We further demonstrate adaptability by retraining the model from an initial sensor to a second capacitive sensor, the Robotiq TSF-85, operating under the same sensing principle. The complete generation pipeline executes in 8.04 ms per frame on the laptop configuration and 6.71 ms on the workstation, enabling real-time operation at 60 Hz on both, and at 120 Hz on the workstation. The similarity of the generated tactile data for the Robotiq TSF-85 is evaluated using complementary similarity metrics, achieving a mean Structural Similarity Index Measure (SSIM) of 0.727 ± 0.16 and a mean Pearson correlation of 0.87 ± 0.16 against real measurements. To demonstrate the utility of the proposed framework, a shape-recognition task (cylinder, sphere, cube) was performed using only synthetic tactile data and evaluated on real-world sensor data, achieving an accuracy of 69.3% with zero real training labels. By enabling integrated tactile simulation and synthetic data generation within Isaac Sim, this work provides a practical tool for tactile perception using capacitive tactile sensors. Full article

This article examines the challenge of trajectory inference for an unknown system affected by external disturbances, with the objective of reconstructing the trajectory of a quadrotor using a reference model. The proposed methodology extends the Closed-Loop Output Error (CLOE) scheme through two complementary contributions: an identified gain, incorporated into the reference model to guarantee the Hurwitz condition of the closed-loop error dynamics, and a set of sliding-mode correction terms that further accelerate error convergence and enhance robustness against bounded disturbances. The stability of both contributions is formally established via Lyapunov-based analysis. The proposed approach is validated through realistic simulations carried out in the CoppeliaSim robotics environment, considering both constant and time-varying trajectory scenarios. Results show that the hybrid approach improves trajectory inference accuracy and convergence speed, maintaining resilience under adverse conditions, making it a promising alternative for autonomous quadrotor monitoring. Full article

Object detection remains a fundamental challenge in computer vision and plays a pivotal role in safety-critical medical applications, including surgical instrument recognition and operating-room workflow automation. This study presents a comprehensive comparative evaluation of five attention mechanisms—Squeeze-and-Excitation (SE), Convolutional Block Attention Module (CBAM), Efficient Channel Attention (ECA), Simple Attention Module (SimAM), and an enhanced multi-kernel Spatial Pyramid Pooling Fast module (SPPF+)—integrated into the YOLOv26n backbone, together with two neck-level attention variants (ECA-Neck and CBAM-Neck). A total of 16 model configurations were systematically investigated on a 22-class dental surgical instrument detection dataset under both default training settings and hyperparameter configurations optimized using the Optuna Tree-structured Parzen Estimator (TPE), enabling a rigorous full-factorial ablation study. Experimental results demonstrate that TPE-based hyperparameter optimization consistently enhances detection performance across all architectures. Among the evaluated models, CBAM-Opt achieved the highest detection accuracy, attaining an mAP@50 of 0.959 and an F1-score of 0.913, although the margins among the top optimized configurations fall within run-to-run variability. In contrast, Base-Opt delivered the strongest strict-localization capability with an mAP@50–95 of 0.800, highlighting the competitive performance of the baseline architecture when appropriately optimized. Notably, the parameter-free SimAM module exhibited the largest improvement following optimization ($\Delta$mAP@50 = +0.040), indicating a pronounced sensitivity to training configuration. Furthermore, neck-level attention integration achieved performance comparable to backbone-based attention, with ECA-Neck-Opt reaching an mAP@50 of 0.959, suggesting an effective alternative that preserves pretrained feature representations while maintaining high detection accuracy. Beyond performance evaluation, this work provides a unified benchmarking framework for attention mechanisms in medical object detection, accompanied by computational complexity analysis and practical architectural insights. The findings establish evidence-based guidelines for selecting attention modules in resource-aware surgical vision systems and contribute toward the development of more accurate and reliable computer-assisted clinical workflows. Full article

Product–Service Systems (PSS) require feedback loops that connect stakeholder experience with design decisions, but distributed feedback is often difficult to transform into actionable design knowledge. This paper synthesises previously distributed results from the DIVERSITY project into an integrated decision-support framework for feedback-driven PSS design. The framework combines social-sentiment representation, reach/influence weighting, opinion modelling, monitoring, extraction and prediction modules, and a hierarchical similarity mechanism for exploring likely sentiment toward new PSS configurations. It was implemented in the DIVERSITY engineering environment and assessed through a preliminary polarity plausibility check, controlled testing with synthetic data generated by OpinionSim, and three industrial business cases. The results indicate that the framework can support segmentation-aware interpretation of stakeholder feedback, shorten the path from market signals to design action, and provide exploratory evidence for early PSS concept generation and redesign. The predictive component is presented as an exploratory design-support mechanism, with forecasting accuracy remaining a topic for future empirical validation. The paper contributes an integrated feedback-to-design framework for PSS and clarifies the conditions under which sentiment-based indicators are most useful in B2C and B2B settings. Full article

Robotics simulation is a key enabler for automation and control development, allowing safer experimentation, faster design iterations, and reduced development cost. Nevertheless, the current simulator ecosystem is highly fragmented, spanning open-source and commercial tools with different levels of physical fidelity, performance, and ecosystem integration. As a result, simulator selection is frequently driven by familiarity or availability rather than by explicit task requirements, often leading to suboptimal engineering workflows. This paper proposes a task-oriented decision framework to support reproducible and transparent selection of robotics simulators based on a fixed and structured set of evaluation criteria. These criteria cover (i) physical fidelity and contact modelling; (ii) sensor modelling and visual realism; (iii) performance and scalability aspects, including headless execution, parallelism, and GPU acceleration; (iv) ecosystem integration with automation, control, and learning pipelines, including ROS/ROS 2 compatibility; (v) extensibility and programmability; and (vi) practical constraints such as hardware requirements, licensing models, and learning curve. The framework is operationalised through a checklist and scoring matrix guided by four key questions addressing the target task, fidelity-versus-speed priorities, target software stack, and sim-to-real transfer requirements. To examine feasibility in a representative engineering workflow, a URDF-based modelling and simulation pipeline is implemented and used to compare Gazebo, as an open-source physics-based simulator, against MATLAB/Simulink, representing a commercial model-based simulation environment. The comparison reports practical indicators including setup effort, integration complexity, computational requirements, and runtime behaviour for repeated executions of a representative motion-oriented sequence. The results highlight consistent trade-offs across different user profiles and application needs while also revealing open gaps in the field, notably the lack of unified multi-task benchmarks and joint metrics capable of simultaneously capturing simulation fidelity, computational performance, and sim-to-real transfer effectiveness. Full article

Large language models (LLMs) exhibit strong general conversational capabilities, yet their deployment in domain-specific service environments such as hospitality remains limited by the absence of emotionally grounded datasets and validated end-to-end system architectures. This paper presents HOSPIT-LLM, an EU-funded euROBIN Technology Exchange Program pilot, as a complete, integrated pilot pipeline for human-centered conversational AI in hotel reception scenarios. We deploy a multimodal hotel-terminal assistant in a real hotel reception, capturing synchronized dual-camera video and audio to collect authentic guest–staff interactions. Speech is transcribed using Whisper, and emotion is extracted from the corresponding video segments via DeepFace, producing 582 real Greek guest–receptionist exchange examples. The resulting data are classified into eight Standard Operating Procedure (SOP) categories. To address data scarcity, we augment the corpus with 1269 synthetic dialogues generated by eight diverse LLMs through the OpenRouter API, yielding a total of 1851 dialogue records with explicit emotion-token annotation. We fine-tune Qwen3.5-35B-A3B using Low-Rank Adaptation (LoRA) through a two-stage process: supervised fine-tuning (SFT) on an 888-example conversation pool and Simple Preference Optimization (SimPO) on a 1899-pair preference pool, each split 80/10/10 into training, validation, and test. The resulting model is integrated into an interactive hotel-terminal system combining YOLO-based person detection, face-recognition-driven guest personalization, Kokoro neural text-to-speech (TTS), and a multi-service orchestration layer connected to the hotel Property Management System (PMS). Evaluation combines standard text metrics, emotion-aware scoring, and a large-model judge. The results indicate targeted improvements in the rule-based contextual emotion-policy match and staff-emotion policy compliance compared to the base model, while general response-quality gains remain more modest. In particular, the rule-based contextual policy-match score improves from 0.614 to 0.901, while forbidden staff-emotion outputs decrease from 0.142 to 0.018. The deployed pilot demonstrates the practical integration of a personalized, emotion-aware LLM assistant in an interactive hotel-terminal setting; end-to-end latency and fully hotel-side edge deployment were not evaluated and are left for future work. HOSPIT-LLM provides a reproducible framework for multimodal dataset creation, preference-based fine-tuning, and deployment of human-centered AI systems. A mobile robotic embodiment is planned as future work. Full article

To address the limited maneuverability and tracking accuracy of autonomous four-wheel steering (4WS) vehicles during parallel parking in confined spaces, an adaptive linear time-varying model predictive control (LTV-MPC) strategy for integrated path tracking and steering coordination is proposed. Different from conventional MPC-based parking controllers with fixed weighting parameters and steering allocation schemes, the proposed method introduces an adaptive weighting mechanism that adjusts the tracking-error weights online according to the yaw-angle error, thereby improving the balance between tracking accuracy and control smoothness throughout the parking process. A hyperbolic tangent (tanh)-based steering allocation strategy is further developed to realize smooth transitions between reverse-phase steering and posture adjustment, while a PID compensation module is incorporated to improve steering command execution. A kinematic single-track model considering Ackermann steering geometry is established to formulate the prediction model of the controller. MATLAB/Simulink and CarSim co-simulation, together with hardware-in-the-loop (HIL) experiments, are conducted to evaluate the proposed control strategy. The experimental results demonstrate that the proposed method achieves more accurate path tracking, smoother steering responses, and higher parking stability than conventional controllers under typical parallel parking scenarios. The proposed strategy provides an effective and practical control framework for improving the low-speed maneuverability and path-tracking performance of autonomous 4WS vehicles. Full article

Unmanned aerial vehicles (UAVs) have become essential cyber-physical platforms for applications such as surveillance, infrastructure inspection, emergency response, and intelligent transportation. However, their tight coupling among sensing, communication, control, actuation, and swarm coordination also exposes them to sophisticated cyber-physical attacks that are difficult to detect using conventional intrusion detection systems. Existing machine learning, deep learning, graph-based, and federated intrusion detection approaches generally rely on statistical feature representations or temporal patterns, providing limited capability to model causal dependencies among interacting UAV subsystems and to generalize across heterogeneous operating environments. To address these limitations, this paper proposes Fed-CGIDS-UAV, a federated causal graph learning framework for cross-domain intrusion detection in cyber-physical UAV networks. The proposed framework models each telemetry window as a typed causal graph in which nodes represent navigation, sensing, communication, control, actuation, and swarm states, while directed edges capture stable operational dependencies. Intrusions are detected by identifying violations of these learned causal relationships, and the framework provides interpretable node-edge explanations to support root-cause analysis. Furthermore, federated learning enables collaborative model training across distributed UAV clients without sharing raw telemetry, thereby preserving data privacy while improving robustness under heterogeneous operating conditions. The proposed framework was implemented and experimentally evaluated in a controlled simulation environment covering four UAV operating domains and six representative attack classes. All experiments were repeated over five independent runs using different random seeds, and the reported results correspond to the measured average performance. The proposed framework was implemented using Python 3.12 (Python Software Foundation, Wilmington, DE, USA) and PyTorch 2.3 (Meta Platforms, Menlo Park, CA, USA). UAV flight data were generated using Microsoft AirSim 1.9.1 (Microsoft Corporation, Redmond, WA, USA), integrated with PX4 Autopilot v1.14 (Dronecode Foundation, San Francisco, CA, USA) and Gazebo Sim 11 (Open Source Robotics Foundation, Mountain View, CA, USA). Within this simulation-based evaluation, Fed-CGIDS-UAV achieved an accuracy of 0.968, an F1-score of 0.956, and an internal–external stability gap (IESG) of 0.028, outperforming conventional machine learning, deep learning, graph-based, and centralized causal baselines while maintaining competitive computational latency. Although these results demonstrate the effectiveness of the proposed framework under controlled simulation conditions, validation using real-flight UAV telemetry remains an important direction for future research. These results demonstrate that integrating causal graph learning with federated optimization provides an effective and interpretable solution for privacy-preserving intrusion detection in heterogeneous cyber-physical UAV environments. Full article