Frontiers in Neurorobotics最新文献

{"title":"A bionic intelligent method combining evolutionary game theory with particle swarm optimization for UAV 3D path planning.","authors":"Lixin Jia, Peng Shi","doi":"10.3389/fnbot.2026.1857152","DOIUrl":"10.3389/fnbot.2026.1857152","url":null,"abstract":"<p><p>Recently, UAV path planning in 3D complex environments has attracted increasing attention due to its significance in UAV motion control systems. However, the NP-hard nature of this problem poses significant challenges in generating a high-quality path. To address this issue, this paper proposes an improved self-adaptive particle swarm optimization (ISAPSO) algorithm by integrating the standard PSO 2011 with evolutionary game theory (EGT). Firstly, a novel self-adaptive parameter updating strategy is proposed, which combines the evolutionary stable strategy in EGT with hyperbolic tangent function to balance the exploration and exploitation capabilities of ISAPSO. Subsequently, an ISAPSO-based path planning approach is developed to generate optimal 3D path for UAV in an obstacle-rich environment. To efficiently handle constraints, a novel self-adaptive constraint handling technology is proposed in the developed path planner. Finally, the performance of the proposed ISAPSO is evaluated against six state-of-the-art evolutionary algorithms using 20 test functions. Following the benchmark study, the ISAPSO-based path planner is is validated in different scenarios against six well-known counterparts. The simulation results confirm that the proposed ISPASO outperforms its competitors in the benchmark study at a 90% confidence level. Moreover, the ISPASO-based path planning method dominates its contenders in terms of the path optimality. Therefore, the proposed method could be regarded as a vital alternative in the area of path planning.</p>","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":"20 ","pages":"1857152"},"PeriodicalIF":3.5,"publicationDate":"2026-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13357615/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148436369","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Confidence-driven adaptive time window for real-time driver fatigue detection in Level 2-3 autonomous vehicles: a multi-dataset validation study.","authors":"Wantong Xie, Peng Xiao","doi":"10.3389/fnbot.2026.1857548","DOIUrl":"10.3389/fnbot.2026.1857548","url":null,"abstract":"<p><p>Driver fatigue constitutes a critical safety hazard in Level 2-3 (L2-3) conditionally automated vehicles, where the paradoxical demand for sustained supervisory vigilance despite minimal active engagement accelerates cognitive underload and impairs timely takeover readiness. Existing vision-based driver monitoring systems are constrained by fixed temporal analysis windows and binary classifiers that neither quantify prediction uncertainty nor adapt to the heterogeneity of real-world fatigue dynamics, resulting in elevated false alarm rates and poor cross-domain generalization. This study introduces a confidence-driven adaptive time window (CDATW) framework: a closed-loop neuro-computational pipeline that couples a lightweight MobileNetV3-CBAM-BiLSTM spatial-temporal encoder with a Monte Carlo Dropout uncertainty estimator to produce simultaneous fatigue probability and epistemic confidence outputs at each inference step. The confidence signal governs a window controller that contracts observation periods to 5-10 s under high certainty (confidence >0.85) for rapid warning, and extends them to 20-30 s under low certainty (confidence <0.60) to suppress spurious alarms-instantiating the feedback-driven adaptive sensing principle central to neurorobotic perception. The framework was validated on four heterogeneous public datasets (NTHU-DDD, YawDD, UTA-RLDD, and DROZY) under single-dataset, cross-dataset transfer, and mixed-dataset training protocols. Single-dataset accuracy ranged from 88.6 to 91.8% with AUC of 0.92-0.95, while the adaptive mechanism reduced false alarm rates by 35.2% relative to fixed 15-s baselines. The architecture sustains 38-45 FPS on an NVIDIA Jetson Xavier NX automotive embedded platform, and confidence calibration achieves an Expected Calibration Error of 0.078, with high-confidence predictions (>0.9) attaining 95.6% accuracy. These results demonstrate that uncertainty-aware adaptive temporal reasoning embedded in a deployable neurorobotic architecture constitutes a computationally efficient and practically viable strategy for driver state monitoring in L2-3 autonomous vehicles, with broader implications for closed-loop perception in safety-critical human-machine systems.</p>","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":"20 ","pages":"1857548"},"PeriodicalIF":3.5,"publicationDate":"2026-06-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13337876/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148404400","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Learning robust and generalizable bimanual skills: a spatiotemporal causal hierarchical diffusion framework with attention anti-interference.","authors":"Xukun Liu, Fengjuan Xie, Zhenyu Liu, Guangning Li, Shenggang Wei, Kai Xu, Aifeng Liu","doi":"10.3389/fnbot.2026.1849143","DOIUrl":"10.3389/fnbot.2026.1849143","url":null,"abstract":"<p><strong>Introduction: </strong>Bimanual visuomotor imitation learning enables robots to acquire coordinated dual-arm manipulation skills from visual demonstrations, yet it faces significant challenges in temporal synchronization, spatial collision avoidance, long-horizon reasoning, and robustness to visual distractions. Existing diffusion-based policies often struggle to simultaneously capture long-horizon temporal dependencies and fine-grained spatial precision, while remaining sensitive to spurious correlations and domain shifts.</p><p><strong>Methods: </strong>To address these limitations, we propose the Spatiotemporal Causal Hierarchical Diffusion Imitation Learner (SCH-DIL), a framework that integrates spatiotemporal hierarchical diffusion optimization to factorize the denoising process into temporal and spatial branches for multi-scale action modeling. The framework further incorporates causal visual representation learning that minimizes mutual information with confounding environmental factors to produce invariant features, along with noise-robust diffusion modeling that employs learnable observation uncertainty estimation and confidence-aware denoising. Additionally, attention anti-interference regularization is introduced to penalize distractions and enforce temporal attention consistency.</p><p><strong>Results: </strong>Extensive experiments on the RoboTwin 2.0 benchmark demonstrate that SCH-DIL consistently outperforms existing diffusion-based and imitation learning baselines, achieving higher success rates under both clean and domain-randomized inference settings.</p><p><strong>Discussion: </strong>These improvements are achieved with minimal computational overhead, suggesting that the proposed hierarchical and causally regularized diffusion framework offers a practical and robust solution for bimanual visuomotor imitation learning in visually challenging and domain-shifted environments.</p>","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":"20 ","pages":"1849143"},"PeriodicalIF":3.5,"publicationDate":"2026-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13333590/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148436303","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DSPE-ViT: a lightweight vision transformer with dynamic sparse positional encoding for dense small object detection in UAV imagery.","authors":"Liya Cai, Shuping Li","doi":"10.3389/fnbot.2026.1849093","DOIUrl":"10.3389/fnbot.2026.1849093","url":null,"abstract":"<p><strong>Background: </strong>Detecting densely distributed small objects in unmanned aerial vehicle (UAV) aerial imagery poses a persistent challenge in computer vision. Vision Transformers (ViTs), empowered by global self-attention, perform strongly in object detection, but their fixed absolute positional encoding (PE) adapts poorly to scenes where small targets cluster at high density, and redundant encoding dimensions introduce unnecessary computational overhead.</p><p><strong>Methods: </strong>This paper presents DSPE-ViT, a lightweight ViT-based detection framework tailored for dense small object detection in UAV imagery. Its core DSPE module comprises two complementary components: a PE Redundancy Pruner that employs learnable soft-gating masks to adaptively suppress redundant PE dimensions, and a Local PE Enhancer that introduces density-aware adaptive-window relative positional encoding to strengthen local spatial perception in high-density regions. Beyond the DSPE module, a Small Object Feature Pyramid Network (SmallObjFPN) integrating SE channel attention with depthwise separable convolutions improves multi-scale feature representation, and the WIoU v3 loss is adopted to refine bounding-box regression for small targets.</p><p><strong>Results: </strong>On the VisDrone2019-DET dataset, DSPE-ViT achieves 43.2% mAP@0.5 with only approximately 6.0 M parameters and 15.8 GFLOPs. Cross-domain evaluation on SeaDronesSee yields 30.1% mAP@0.5 under zero-shot transfer and 38.4% after fine-tuning.</p><p><strong>Conclusion: </strong>The cross-domain results confirm the generalization capability of the proposed lightweight framework.</p>","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":"20 ","pages":"1849093"},"PeriodicalIF":3.5,"publicationDate":"2026-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13316741/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148360026","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ST-HONet: Spatio-Temporal Hierarchical Network for long-horizon bimanual visuomotor imitation.","authors":"Xukun Liu, Fengjuan Xie, Kai Xu, Zhenyu Liu, Shenggang Wei, Guangning Li, Xu Sun","doi":"10.3389/fnbot.2026.1860170","DOIUrl":"10.3389/fnbot.2026.1860170","url":null,"abstract":"<p><strong>Introduction: </strong>Learning robust and temporally consistent manipulation policies from long-horizon visual observations remains a fundamental challenge in imitation learning. While recent Transformer-based approaches reduce compounding errors via temporally extended action chunks, most methods rely on deterministic or unimodal action representations, limiting their ability to capture the inherent multimodality of expert demonstrations.</p><p><strong>Methods: </strong>In this work, we propose ST-HONet, a Spatio-Temporal Hierarchical Network that integrates three adaptive optimization mechanisms operating jointly across spatial and temporal dimensions. ST-HONet formulates policy learning as conditional generation of extended action segments using a Transformer-based conditional variational autoencoder to explicitly model multimodal expert behaviors through a structured latent representation. To ensure stable optimization over long temporal horizons, we introduce a unified spatio-temporal training framework combining adaptive data augmentation, progressive latent regularization, and multi-stage optimization strategies.</p><p><strong>Results: </strong>We evaluate ST-HONet on RoboTwin 2.0, a large-scale benchmark for long-horizon bimanual manipulation under domain randomization with automatically generated expert demonstrations. Across representative manipulation tasks, ST-HONet achieves consistently higher task success rates compared to baseline models, while incurring only minimal additional computational overhead.</p><p><strong>Discussion: </strong>These results demonstrate that explicitly modeling multimodality via a structured latent space, combined with joint spatio-temporal training mechanisms, significantly improves policy robustness and temporal consistency in long-horizon visuomotor imitation learning. The minimal computational overhead further supports the practical deployability of ST-HONet in real-world systems.</p>","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":"20 ","pages":"1860170"},"PeriodicalIF":3.5,"publicationDate":"2026-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13303966/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148344855","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exoskeleton robotics: from rigid structures to bio-integrated systems.","authors":"Liancheng Zheng, Rizauddin Ramli, Wenfeng Zhang","doi":"10.3389/fnbot.2026.1850367","DOIUrl":"10.3389/fnbot.2026.1850367","url":null,"abstract":"<p><p>Exoskeleton robots have become a representative class of wearable robotic systems for rehabilitation, mobility assistance, occupational support, and human performance augmentation. As the field moves from laboratory prototypes toward clinical, industrial, and daily life deployment, research priorities are shifting from device-centered performance improvement to human-centered integration. This mini review provides a structured and critically oriented synthesis of exoskeleton technologies from four interconnected perspectives: technical architecture, technological paradigm evolution, deployment barriers, and future research directions. To improve transparency and reproducibility, we adopted a narrative review strategy with explicit literature selection criteria. Publications were identified from major scientific databases using combinations of keywords related to exoskeleton robotics, actuation, control, human-robot interaction, soft robotics, neural interfaces, rehabilitation, and wearable assistance. Representative studies were selected according to relevance, technical influence, clinical or engineering significance, and coverage of major technological paradigms. The review first analyzes three core technical dimensions-actuation systems, control strategies, and human-robot interaction which jointly determine the performance, adaptability, and usability of exoskeleton systems. Rather than only summarizing these technologies, we compare their trade-offs in terms of power density, control precision, compliance, energy efficiency, personalization, safety, and deployment readiness. The review then examines the evolution from rigid exoskeletons, which provide high structural support and precise force transmission, to soft exoskeletons, which improve compliance and comfort, and further to bio-integrated systems that combine neural interfaces, functional electrical stimulation, multimodal sensing, and mechanical assistance. Based on this synthesis, we organize the review using a Human-Exoskeleton Integration Maturity Framework spanning mechanical coupling, physical compliance, functional adaptation, and cognitive/bio-integrated coupling. Persistent barriers, including energy supply, personalization, safety assurance, cost, regulatory translation, and ethical governance, are critically discussed. Finally, future directions are outlined, including neural-interface-driven control, multimodal perception, human-in-the-loop optimization, hybrid rigid-soft architectures, and socially responsible design. Overall, this review argues that the next stage of exoskeleton development will depend not merely on stronger actuators or more intelligent algorithms, but on integrated systems that are adaptive, trustworthy, affordable, and seamlessly embedded in human movement and function.</p>","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":"20 ","pages":"1850367"},"PeriodicalIF":3.5,"publicationDate":"2026-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13351094/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148420489","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ST-HADP: Spatio-Temporal hierarchical attention diffusion policy for long-horizon generalizable bimanual visuomotor imitation.","authors":"Xukun Liu, Fengjuan Xie, Shibo Liu, Xu Sun, Zhenyu Liu, Guangning Li, Shenggang Wei","doi":"10.3389/fnbot.2026.1846433","DOIUrl":"10.3389/fnbot.2026.1846433","url":null,"abstract":"<p><strong>Introduction: </strong>Dual-arm robotic manipulation presents fundamental challenges in coordinating spatially shared perception and temporally extended behaviors under limited demonstration settings. Existing diffusion-based visuomotor policies rely on flat temporal horizons and globally pooled visual features, which fail to capture the structured nature of bimanual collaboration.</p><p><strong>Methods: </strong>We propose the Spatio-Temporal Hierarchical Attention Diffusion Policy (ST-HADP), a framework that extends 3D diffusion policies through explicit spatial and temporal structuring. ST-HADP introduces a Spatial Attention Module that learns arm-specific focus over task-relevant 3D regions, enabling dynamic and coordinated spatial reasoning. It further incorporates a Temporal Abstraction Module that models action sequences across multiple timescales via hierarchical latent variables, facilitating coarse-to-fine action generation aligned with the natural progression of long-horizon tasks. These components are jointly optimized with a multi-objective loss function that integrates attention regularization and temporal consistency, promoting spatially focused and temporally smooth coordination.</p><p><strong>Results: </strong>We evaluate ST-HADP on the RoboTwin 2.0 platform across six dual-arm robot configurations with diverse morphologies and tasks. Using only 50 automatically generated expert demonstrations, our method consistently outperforms baseline policies, achieving higher success rates with modest additional computational overhead.</p><p><strong>Discussion: </strong>The results demonstrate that explicit spatial and temporal structuring enables effective dual-arm coordination under limited demonstration settings. ST-HADP provides a generalizable framework for bimanual manipulation, suggesting that hierarchical attention mechanisms offer a promising direction for sample-efficient learning of coordinated multi-arm behaviors.</p>","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":"20 ","pages":"1846433"},"PeriodicalIF":3.5,"publicationDate":"2026-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13283968/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148307877","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"EQISP: efficient quantized image signal processing with multi-scale pyramid fusion for resource constrained embodied perception.","authors":"Tongxin Yang, Ling Guo, Jie Li, Qi Qin, Bowen Liu, Qian Zhang","doi":"10.3389/fnbot.2026.1866481","DOIUrl":"10.3389/fnbot.2026.1866481","url":null,"abstract":"<p><strong>Introduction: </strong>Resource-constrained environmental perception requires autonomous robots and embodied intelligent systems to process visual signals efficiently while preserving image fidelity in complex real-world environments. However, converting high dynamic range RAW sensor data into perceptually faithful RGB images remains computationally expensive, thereby limiting the deployment of neural image signal processors on edge platforms with restricted memory, energy, and computational budgets.</p><p><strong>Methods: </strong>Consequently, this study proposes the enhanced quantized image signal processor (EQISP), comprising the quantized convolutional neural network (QCNN) and the unified pyramid fusion algorithm (UPFA). QCNN employs dynamic fixed-point hybrid quantization, which adjusts parameter ranges according to the linear relationship between threshold standard deviation and fractional length, thereby significantly reducing the computational load. Meanwhile, UPFA utilizes Gaussian pyramids to capture global illumination and Laplacian pyramids to preserve fine details, enabling multi-scale, multi-exposure fusion and iterative reconstruction to mitigate detail loss induced by quantization.</p><p><strong>Results: </strong>Comprehensive comparative experiments demonstrated that EQISP achieved a PSNR of 22.90 dB, an SSIM of 0.9278, and 164.843 GFLOPs. Compared with the PyNET baseline, EQISP improved the PSNR by 1.71 dB while reducing the computational cost by a factor of 4.24. Furthermore, deployment experiments on an NVIDIA Jetson TX2 development board showed that EQISP achieved a model size of 57 MB, an inference latency of 189 ms, an inference speed of 6.1 FPS, and a peak memory usage of 2.2 GB.</p><p><strong>Discussion: </strong>These results provide practical evidence that EQISP can serve as an efficient and scalable visual front end for resource-constrained embodied perception systems.</p>","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":"20 ","pages":"1866481"},"PeriodicalIF":3.5,"publicationDate":"2026-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13284068/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148307907","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Research on embodied agent multimodal perception and real-time path planning algorithms for complex unstructured environments.","authors":"Hexuan Ren","doi":"10.3389/fnbot.2026.1846108","DOIUrl":"10.3389/fnbot.2026.1846108","url":null,"abstract":"<p><p>Autonomous navigation of embodied agents in complex unstructured environments demands tightly coupled multimodal perception and real-time path planning capabilities, forming a core technical bottleneck in physical-world robot deployment. Heterogeneous sensor data from visual, LiDAR, and depth modalities remain difficult to align and fuse under varying illumination and terrain conditions, while dynamic obstacle configurations impose severe latency constraints that existing planning algorithms fail to satisfy simultaneously. This paper proposes an integrated end-to-end framework combining a Cross-Modal Attention Fusion (CMAF) module, a Kalman-Graph Neural Network (K-GNN) dynamic obstacle predictor, and a two-layer Proximal Policy Optimization path planning architecture. The Cross-Modal Attention Fusion module fuses three-modal features through a multi-head attention mechanism, achieving a mean Intersection over Union of 78.6% with a fusion latency of 5.3 ms on a self-built unstructured environment dataset. The Kalman-Graph Neural Network couples Kalman filter physical motion priors with graph neural network interaction modeling to predict short-term trajectories of multiple moving obstacles online. The two-layer planner integrates fused perception features with a global semantic topology path to output local velocity commands in real time, reducing average planning time to 18.4 ms. Experiments on a Gazebo simulation platform and a self-developed four-wheeled robot across 60 unstructured test cases demonstrate a navigation success rate of 94.5%, surpassing the strongest baseline by 7.8 percentage points and satisfying real-time operational requirements.</p>","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":"20 ","pages":"1846108"},"PeriodicalIF":3.5,"publicationDate":"2026-06-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13272145/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148277163","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"NL-YOLOv5: a model with a larger receptive field and the ability to globally acquire features.","authors":"Zhiyu Li, Jinhu Liu, Zhihao Zhuo, Lin Chen, Xiao Zeng, Di Li, Yongfa Zhou, Chunzhou Huang","doi":"10.3389/fnbot.2026.1764856","DOIUrl":"10.3389/fnbot.2026.1764856","url":null,"abstract":"<p><strong>Introduction: </strong>Landslide disasters cause severe casualties and economic losses, demanding rapid and accurate detection from high-resolution remote sensing imagery. Traditional methods struggle with insufficient landslide samples and low detection accuracy.</p><p><strong>Methods: </strong>To address this, we propose a dual solution: (1) DCGAN-based data augmentation generating 3,429 synthetic landslide samples from Google Earth imagery, significantly improving model generalization; (2) an NL-YOLOv5 model integrating non-local attention (NLA) and an improved LK-SPP module (based on large-kernel convolution concepts) to enhance global information capture.</p><p><strong>Results: </strong>The NL-YOLOv5 model achieves 80% detection accuracy (a 5% improvement over baseline), with 7% higher precision, 2% higher recall, and 5% higher F1-score, while maintaining real-time speed at 69 f/s.</p><p><strong>Discussion: </strong>This work delivers a practical solution for high-precision landslide detection in remote sensing applications.</p>","PeriodicalId":12628,"journal":{"name":"Frontiers in Neurorobotics","volume":"20 ","pages":"1764856"},"PeriodicalIF":3.5,"publicationDate":"2026-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13260476/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148251041","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}