Search Results (701)

Line-driven soft robots possess inherent advantages in terms of cushioning and terrain adaptability, but the controllable deformation design of line-driven structures and its coordination mechanism with the overall robot motion remain insufficiently studied. To fill this gap, this paper designs a line-driven hexapod soft robot that achieves directional bending of flexible legs through unilateral line traction, combined with triangular gait co-motion and ROS-based multi-sensor perception. Integrating leg deformation as part of the motion mechanism enables the robot to achieve straight-line and turning movements while maintaining structural compliance. This paper establishes the mapping relationship between the leg actuation space, configuration space, and task space, constructs a kinematic model, and uses the finite element method to analyze leg deformation and stress distribution. Based on this, a robot prototype is built, and a ROS-based distributed control and perception system is constructed, utilizing LiDAR, camera, and attitude sensor data to achieve SLAM and state monitoring. Experimental results show that the robot can achieve continuous motion with an average speed of 15.32 mm/s and a turning angle of 4.75° in a single gait cycle. The feasibility of line-driven structure control based on unilateral traction was verified, and a reference was provided for the design of soft robots oriented towards environmental perception. Full article

Continuous full-domain vehicle trajectories are essential for smart highway monitoring, but single-sensor roadside perception is limited by physical coverage, occlusion, and environmental sensitivity. To address continuous trajectory tracking across multiple roadside-sensing domains, this study proposes a real-time, full-domain highway trajectory tracking framework based on radar–camera fusion, improved Deep-SORT, and inverse covariance intersection. At the local perception level, a two-stage object-level and decision-level fusion model is constructed, and Deep-SORT is improved using a CIoU matching strategy and an occluded target tracking controller to enhance local multi-object tracking continuity. At the cross-domain association level, a geometry-motion consistency stepwise calibration method is developed to unify adjacent sensing domains, and a CATS-ICI trajectory stitching strategy is introduced to improve trajectory association and state smoothness during sensor handover. The proposed framework was validated on a real highway test section with roadside radar, video, and drone-based ground-truth trajectories. Experimental results show that the full local method achieves an EMOTA of 92.35%, and the reconstructed full-domain trajectories achieve a successful trajectory matching rate of 98.4% under the 452 vehicles/10 min test condition. Additional ablation experiments further verify the contributions of radar–camera fusion, CIoU, OTTC, GMCSC, CATS, and ICI. These results demonstrate that the proposed framework can provide continuous and reliable full-domain vehicle trajectories for real-world highway monitoring. Full article

To achieve non-contact and continuous monitoring of body length in Micropterus salmoides and overcome the stress damage and subjective error associated with traditional manual measurement, this paper proposes an improved YOLOv8-based multi-target tracking framework for intensive recirculating aquaculture systems. The system employs a geometric measurement framework based on monocular vision that achieves conversion from pixel coordinates to actual body length through camera calibration, water-surface refraction correction, and pose projection correction. Under a collaborative optimization framework integrating detection and tracking, the model incorporates multi-scale feature enhancement, lightweight re-identification (ReID), and a robust data association mechanism, which improves system stability under conditions of high fish density, variable illumination, and turbid water. A shallow feature fusion path is introduced to enhance small-target perception, and a MobileNetV3_ReID model is adopted to extract highly discriminative appearance features, which improves identity consistency while maintaining model compactness. In the data association stage, a hybrid cost matrix integrating IoU, cosine similarity, and motion consistency is constructed, and optimal matching is realized through the Hungarian algorithm. Dynamic threshold adjustment and an exponential moving-average feature-update strategy are introduced to effectively suppress identity switching. Experiments were conducted on an overhead video dataset of Micropterus salmoides collected at a recirculating aquaculture system facility. The results show that the proposed method achieves 82.7% mAP50 while maintaining a real-time throughput of 88 FPS, with MOTA reaching 76.9% and IDF1 reaching 81.5%—the latter representing an improvement of 3.2 percentage points over BoT-SORT and 5.3 percentage points over the YOLOv8 baseline tracker. The number of identity switches (IDSW) decreased from 89 in the baseline configuration to 39, a reduction of 56.2%. Crucially, these component-level improvements translate into a body length error (BLE) of 5.2 ± 1.8% (MAE = 1.35 cm, Pearson r = 0.972), representing a 38.8% improvement over the baseline BLE of 8.5% and satisfying the 5–10% tolerance required for aquaculture growth monitoring. Ablation analysis confirms that both detection enhancements (contributing −1.3% BLE) and tracking optimizations (contributing −2.0% BLE) are necessary to achieve this application-level accuracy. Full article

Deep learning perception in autonomous vehicles (AVs) has created a critical attack surface in which adversarial patches and sensor-spoofing perturbations cascade from perception errors into unsafe driving decisions. Existing defenses face three limitations: most require retraining the perception network, making them impractical for already-deployed fleets; they operate almost exclusively at the perception layer, without verifying whether a compromised detection actually altered the driving action; and they leave temporal consistency across frames largely unexploited. This paper presents SENTINEL, a zero-modification, plug-and-play defense that wraps any deployed AV perception-and-planning stack without updating its weights, calibrating only the detection thresholds, score combination weights, and reference exemplars once on a small held-out calibration set. SENTINEL integrates a frozen foundation model verification ensemble (CLIP, DINOv2, SAM-2), a temporal consistency scorer that flags patches through anomalous frame-to-frame stability under ego-motion, a counterfactual policy verifier that replans under reconstructed perception and measures action-space divergence, and a risk-adaptive safety shield that modulates driving aggressiveness by verification confidence. Across CARLA, nuScenes, KITTI, and BDD100K, against five adversarial attacks and an adaptive adversary, SENTINEL reduces the attack success rate by up to 92%, keeps the clean accuracy loss to approximately 1.8 percentage points, reduces the collision rate under attack by approximately 87%, and adds under 45 ms latency on an RTX 4090 GPU. SENTINEL reframes adversarial robustness as a runtime property of the complete autonomous decision pipeline. Full article

Mobile inspection robots require reliable license plate detection and recognition under constrained computing resources, small-scale or distant imaging conditions, motion blur, and complex background interference. To address these coupled challenges, this paper proposes a lightweight robot-view visual sensing framework for CPU-oriented license plate perception. Instead of simply stacking network modules, the proposed framework follows a unified design principle of reducing redundant computation while compensating for task-critical visual information. In the detection stage, a YOLOv8-MGL detector is developed based on YOLOv8n by combining GhostC2f-based lightweight feature aggregation with LSKAlite-based contextual enhancement after the SPPF module. In the recognition stage, SimAM is embedded into LPRNet to enhance discriminative character responses under motion blur, low resolution, and local degradation without introducing additional learnable parameters. Experiments on the held-out EDRV-LP test set show that YOLOv8-MGL achieves 99.5% mAP₅₀ and 71.1% mAP_50:95, while reducing the number of parameters from 3.01 M to 2.77 M and GFLOPs from 8.1 to 7.5 compared with YOLOv8n. On a CPU-only Intel Xeon Platinum 8260C platform, YOLOv8-MGL achieves 23.98 FPS. SimAM-LPRNet improves the module-level cropped-plate recognition accuracy from 83.10% to 87.17%. To further examine system-level feasibility, a supplementary YOLOv8-MGL + CRNN-CTC pipeline is evaluated from raw images to final plate strings, achieving 91.0% exact recognition accuracy on the held-out EDRV-LP test set, 92.0% on a non-overlapping external CCPD subset, and 13.25 FPS for complete CPU-only processing. These results demonstrate that the proposed framework provides a favorable trade-off among model compactness, localization quality, recognition robustness, and CPU-oriented inference feasibility for mobile robotic inspection scenarios. Full article

Visual perception is an important component of close-range maritime situational awareness, particularly when conventional sources such as AIS and radar are delayed, incomplete, or unavailable. Ship wakes provide motion-related visual cues, but their segmentation remains difficult because wake regions are elongated, weakly textured, and frequently mixed with water-surface clutter. This study develops Shipwake-YOLO, a wake-oriented adaptation of YOLOv9-Seg for ship and wake instance segmentation in inland-waterway images. The task is formulated as visual navigational-state cue extraction rather than validated future manoeuvre prediction. The model segments hull and wake instances and provides mask-derived spatial cues for possible downstream state interpretation. The architecture introduces iAFF into cross-scale feature fusion, adapts the high-level SPPELAN aggregation block with SAConv-enhanced convolution, replaces selected downsampling paths with iSACADown, and adopts MPD-IoU as the bounding-box regression loss. On a 2100-image dataset collected from the Wuhu Channel of the Yangtze River, Shipwake-YOLO improves Box-mAP@50 from 77.7% to 84.6% and Mask-mAP@50 from 67.6% to 79.8% relative to the YOLOv9-Seg baseline. Under stricter IoU thresholds, the model reaches 48.9 in Box-mAP@[0.50:0.95] and 46.2 in Mask-mAP@[0.50:0.95]. The parameter count is reduced by 7.5%, and GFLOPs decrease from 144.2 to 137.1. These results indicate that the proposed adaptation improves ship-wake perception within the collected inland-waterway setting and provides a visual basis for downstream navigational-state estimation. Full article

Virtual reality (VR) provides a flexible platform for investigating human perception of augmented bodily abilities. While motion scaling of virtual limbs has been explored in previous studies, psychophysical detection thresholds and embodiment have rarely been examined together, and direct comparisons between motion enlargement and reduction remain limited. In this study, we systematically manipulated the motion gain of a virtual hand in an immersive VR environment and evaluated both detection thresholds for deviations from unity gain and embodiment, including the sense of ownership and agency. Fifteen participants took part in the experiment. The results showed that the 50% detection thresholds for motion gain were 1.18 for motion expansion and 0.86 for motion shrinkage. These detection thresholds were approximately symmetric about the unity gain. In contrast, embodiment ratings showed an asymmetric decline, with ownership and agency decreasing more steeply for motion shrinkage than for motion expansion. Perceptual detection and subjective embodiment therefore exhibited both consistency and divergence: the onset of significant degradation in ownership and agency occurred near the perceptual detection range, whereas beyond these thresholds, embodiment declined asymmetrically. These findings provide quantitative guidance for designing motion-scaled body augmentation in VR, highlighting the importance of considering not only detectability but also the direction-dependent robustness of embodiment. Full article

Reliable selective picking of long-staple cotton remains challenging because dense dwarf canopies restrict robot operating space and increase boll occlusion, resulting in reduced target visibility and potential fiber damage during picking. To address these challenges, a mobile dual-arm robotic picking system integrating hierarchical depth perception, cotton-boll recognition, optimized motion planning, and three-finger flexible end-effectors was developed for autonomous picking in Xinjiang long-staple cotton fields. The proposed YOLOv7-DCN-SENet model reached 95.75% precision, 92.65% recall, and 97.19% mAP@0.5 on the test set, while the onboard computing platform operated at 101 FPS under the experimental configuration. Indoor and field experiments were conducted on directly visible upper-canopy open cotton bolls. The dual-arm robot achieved parallel picking success rates of 74.6% and 57.6%, with average picking cycles of 28.2 s and 34.9 s, respectively. Field performance was mainly limited by strong-light overexposure, depth-information loss, occlusion-induced localization errors, arm interference within narrow canopy spaces, and incomplete fiber separation during boll detachment. These results demonstrate the feasibility of autonomous dual-arm selective picking for long-staple cotton under dense planting conditions and provide a basis for further improvements in robotic cotton-picking systems. Full article

Background: The vestibular system has traditionally been interpreted within a reflex-based framework, mainly centered on gaze stabilization, vestibulo-ocular reflex pathways, and peripheral vestibular deficits. This model remains essential, but it is insufficient to explain the full spectrum of postural, perceptual, visual-motion, and self-motion complaints observed in contemporary clinical otoneurology. Objective: This conceptual narrative review examines selected representative works by Kathleen E. Cullen as landmarks in a broader transition from reflex physiology to predictive, multimodal, context-dependent, body-centered self-motion control. Methods: This is not a systematic or bibliometric review. Papers were selected because they mark distinct conceptual steps in Cullen’s work: neural encoding of self-motion, peripheral and central coding strategies, multimodal integration, active versus passive self-motion, reafference suppression, body-centered encoding, proprioceptive prediction, vestibular cerebellar internal models, sensory reweighting, and clinical translation. Synthesis: Angelaki and Cullen’s 2008 synthesis and Cullen’s subsequent work demonstrate that vestibular processing is inherently multimodal from the earliest central stages and that neural representations of self-motion depend on behavioral context. Vestibular nuclei, visual-vestibular networks, and vestibular cerebellar circuits integrate labyrinthine signals with optic flow, proprioceptive, oculomotor, motor, cerebellar, cortical, and contextual information. This architecture enables the brain to distinguish expected from unexpected motion, suppress predictable vestibular reafference during voluntary action, compute internal estimates of body motion, adapt to altered sensory reliability, and reweight sensory inputs according to task demands. Conclusions: The clinical relevance of this trajectory is substantial. Patients may show preserved high-acceleration vestibulo-ocular reflex responses while experiencing persistent instability, visually induced dizziness, defective self-motion perception, or abnormal sustained vestibular processing. Such dissociations are not paradoxical when the vestibular system is understood as a predictive, distributed, body-centered control system. Cullen’s long lesson offers a neurophysiological foundation for a modern vestibular grammar in which clinical findings are interpreted across the reflexive, perceptual, postural, visual-vestibular, sustained, and predictive domains. Full article

Low-latency visual perception is essential for interactive machine vision on edge AI devices, but conventional frame-based image sensors impose frame period delays and generate dense image data that increase memory bandwidth and processing latency. Although Dynamic Vision Sensors (DVSs) are known to provide low latency, sparse output, and high dynamic range, these sensor-level properties do not automatically translate into practical application-level latency reduction on resource-constrained edge platforms. This paper presents a latency-driven sensing algorithm co-design approach for DVS-based low-latency machine vision. The main objective is to connect DVS sensor-level characteristics, event representations, task-dependent processing flows, and measured response times on mobile application processors. We first analyze latency requirements for three representative edge AI applications (i.e., person detection, gesture recognition, and Simultaneous Localization and Mapping (SLAM)), which correspond to different latency regimes and processing structures. We then describe the DVS operating principle, pixel-level event latency, and readout latency, showing how asynchronous event generation reduces sensing delay and suppresses redundant static background information before algorithmic processing. In contrast to prior event camera studies that mainly optimize a single task or a specific event representation, this work evaluates three task-specific event processing systems on mobile processors. Person detection achieves 92 ms processing latency on Exynos 7570, gesture recognition based on event-driven 4-DoF motion estimation achieves 20 ms latency on Exynos 5422, and SLAM achieves 15.9 ms latency on Snapdragon 845. These results satisfy the practical latency targets of the corresponding applications and demonstrate that DVS-based sensing can provide not only sensor-level speed advantages but also system-level latency benefits for AIoT, mobile, robotics, and AR/VR machine vision systems. Full article

Background: Virtual reality (VR) has attracted increasing attention in architectural design education because of its potential to support spatial cognition and embodied understanding of architectural space. Compared with conventional two-dimensional (2D) drawings and screen-based three-dimensional (3D CAD) tools, VR enables learners to experience space at a realistic scale through binocular disparity and motion parallax, which may reduce cognitive load and facilitate experiential learning. However, previous studies have mainly relied on subjective evaluations, such as questionnaires and observations, and have not sufficiently examined differences in educational effectiveness among design tools or among students with different learning levels. Objective and Methods: This study aimed to identify effective teaching tools for facilitating students’ understanding at different learning levels and to propose appropriate methods for applying VR to improve educational effectiveness. To achieve this, we proposed an objective experimental method for evaluating the effectiveness of VR in architectural design education based on students’ ability to identify incorrect content in architectural design plans. The experiment compared the performance of students using 2D drawings, 3D CAD, and VR environments and examined differences according to student grade levels (higher- and lower-year students) objectively. Results: The results revealed that both higher- and lower-year students identified more incorrect content items related to “Fitting” (such as door layouts) when using 2D drawings (finding rates were 43.8%~53.3% higher than those with 3D CAD or VR), whereas more incorrect content items related to “Furniture” size were identified when using VR (finding rates were 18.8%~56.3% higher than those with 2D drawings or 3D CAD). In addition, items related to sectional and elevation design, such as “Opening,” as well as issues concerning the size of “Space,” were identified by higher-year students regardless of the tool used. In contrast, lower-year students identified approximately twice as many of these items when using VR as when using 2D drawings. Conclusions: Based on the above results, the effectiveness of VR varied depending on both the type of design knowledge and the students’ learning levels. VR improved lower-year students’ understanding of spatial dimensions, furniture and fitting compared with conventional tools. Furthermore, VR encouraged more detailed consideration of spatial and design-related issues during architectural design tasks. These findings suggest that VR can reduce the cognitive load associated with learning architectural spatial concepts and promote experiential learning close to real spatial perception. Implications: This study supports the appropriate use of VR in architectural design education. The experimental method proposed in this study can also be used to objectively evaluate the effectiveness of educational tools other than VR before their implementation in architectural design education. Applying this method in architectural education is expected to enhance students’ awareness of architectural spatial issues and promote more comprehensive spatial understanding during the design process. Full article

Behavior change technologies are increasingly deployed in everyday contexts where perception errors are difficult to avoid. Such errors can undermine user trust and long-term engagement, while purely technical approaches to error elimination are often impractical in open-world environments. This study proposes a fault-tolerant design that translates algorithmic uncertainty into anthropomorphic expressions of vulnerability. By decoupling task-outcome feedback from internal confidence states, an embodied agent communicates uncertainty through a five-level nonverbal framework comprising posture, facial expression, and motion intensity. The approach was implemented in an interactive waste-sorting system and examined through a three-week field study in a semi-public university corridor. Three feedback strategies were compared: an outcome-only baseline, a persistently confident agent, and an adaptive agent whose vulnerability expression varied according to a transformed confidence signal. The findings suggest differences in user behavior across conditions. Under the adaptive condition, user sorting accuracy exhibited a fluctuation–recovery pattern during the final deployment phase, whereas accuracy under the confident-agent condition showed a declining trend. Correct-trial stay durations were shorter under the adaptive condition, consistent with the formation of a more streamlined interaction routine. In contrast, observations from error cases were limited by the small number of misclassification events. Due to the exploratory nature of the study, the small sample size of the questionnaire and the sequential deployment structure, the results should be considered preliminary evidence. Nevertheless, the findings suggest that expressing vulnerability in an anthropomorphic way may be a promising approach for communicating uncertainty in behavior change systems. Full article

The stroboscopic effect and phantom array effect caused by temporal light modulation (TLM) in light-emitting diode (LED) lighting are important temporal light artifacts (TLAs) that can influence visual perception, task performance, and visual comfort. This review systematically analyzes 40 studies published between 1998 and 2024 to provide a comprehensive overview of the current understanding of both effects. The reviewed literature covers visibility thresholds, influencing parameters, experimental methodologies, and assessment metrics. The analysis shows that reported visibility thresholds for the stroboscopic effect typically range from 550 to 1000 Hz, whereas thresholds for the phantom array effect may extend to 10–15 kHz, suggesting substantial differences in the underlying perceptual mechanisms. In addition to modulation frequency, modulation depth, waveform, duty cycle, luminance, retinal image motion, and observer factors have been identified as important determinants of visibility. The review further highlights significant methodological differences among studies, including variations in experimental design, stimulus generation, participant characteristics, and psychophysical procedures. Although the stroboscopic visibility measure (SVM) provides a standardized framework for evaluating the stroboscopic effect, no comparably validated metric is currently available for the phantom array effect. The review identifies major knowledge gaps regarding the interaction of influencing parameters and the lack of standardized assessment methods. Future research should focus on establishing unified experimental protocols and developing robust metrics for the phantom array effect to support comprehensive lighting standards that protect visual comfort, well-being, and consumer health. Full article

Background/Objectives: Short-form motion graphics are increasingly used in digital health communication. However, limited research has examined their accessibility, interaction quality, and usability for older adults. This study explores the design and evaluation of short-form motion graphics as human–computer interaction systems for depression-related health communication in aging populations. Methods: A data-informed and human-centered approach was adopted, integrating clustering-based analysis, expert evaluation, and user-based assessment. Short-form motion graphics videos and user-generated comments were analyzed to identify design-related themes associated with accessible digital health communication. These insights informed the development of motion graphics prototypes. The evaluation involved independent expert groups and 200 older adult participants. Cognitive load and usability were assessed using a structured questionnaire. Results: The clustering analysis showed moderate cluster separation and provided an exploratory source of design insights. Expert evaluation highlighted the importance of visual clarity, structured content organization, and appropriate motion pacing. User evaluation yielded a mean usability score of 3.95 and a mean cognitive load score of 3.72, indicating generally positive perceptions of the developed motion graphics among participants. Conclusions: The findings suggest that combining data-informed analysis, expert review, and user evaluation may be useful for designing and assessing digital health communication systems for older adults. As this study was exploratory and did not include a control group, the findings should be interpreted within the context of the study and should not be considered evidence of causal relationships. Full article

Facility agricultural robots operating in greenhouse environments often encounter narrow passages, dense obstacle distributions, and frequent path-direction changes, which increase the difficulty of achieving efficient and smooth autonomous navigation. Conventional A* algorithms usually suffer from redundant node expansion, dense turning-point distributions, and insufficient path continuity under such constrained conditions. To address these issues, this study proposes an improved A*-based path-planning framework that integrates adaptive heuristic weighting, dynamic corner correction, and Bézier-curve-based path smoothing. Rather than introducing an entirely new planning paradigm, the proposed method coordinates several existing optimization strategies within a unified framework to improve search efficiency, path regularity, and path continuity for facility agricultural scenarios. The adaptive heuristic weighting strategy dynamically adjusts the contribution of the heuristic term according to the relative distance between the current node and the target node, thereby improving global search guidance while reducing unnecessary exploration. Dynamic corner correction is introduced to suppress zigzag path structures and reduce redundant turning nodes in obstacle-dense regions, while Bézier-curve-based smoothing is employed to improve path continuity and compatibility with the kinematic characteristics of agricultural mobile robots. Simulation experiments were conducted on grid maps and greenhouse-like environments with different obstacle distributions, and comparative evaluations were performed against Dijkstra, RRT, and conventional A* algorithms. Under representative simulation scenarios, the proposed framework reduced the number of turning points by up to 53.7% and decreased computation time by approximately 19.4% compared with the conventional A* algorithm, based on the average results of repeated trials under identical conditions. In addition, physical platform experiments on a ROS2-based agricultural robot demonstrated that the planned trajectories maintained relatively stable navigation performance and smoother directional transitions in constrained greenhouse-like environments. The results indicate that the proposed framework achieves a more balanced trade-off between computational efficiency, path compactness, and path smoothness than the benchmark methods considered in this study. Nevertheless, the current validation remains limited to structured or semi-structured greenhouse environments under static obstacle conditions. Future work will focus on improving adaptability to dynamic agricultural scenarios and integrating the framework with real-time perception and motion-control systems for practical greenhouse deployment. Full article