Search Results (323)

Medication dispensing errors remain a significant concern in healthcare systems, particularly in elderly care and long-term medication management, where incorrect medication delivery may compromise patient safety and treatment outcomes. This study presents the design and experimental validation of a cyber–physical medication dispensing platform integrating robotic manipulation, edge AI-based visual verification, distributed motion control, and cloud synchronization. The platform combines a rotary medication storage mechanism, vacuum-based pill handling, a Klipper-based control framework, and a YOLOv8 perception subsystem deployed on a Hailo AI accelerator for real-time edge inference. Experimental evaluation was conducted under controlled laboratory conditions. Using an environment-specific validation dataset, the perception subsystem achieved a precision of 0.627, recall of 0.739, and mAP@0.5 of 0.786. An adaptive verification strategy was subsequently evaluated to improve dispensing verification under varying pill occupancy conditions. End-to-end system testing comprising 80 dispensing trials achieved an overall dispensing success rate of 86.25%, with no incorrect dispensing events observed. The results demonstrate the feasibility of integrating edge AI verification, distributed control, and cloud connectivity within a cyber–physical medication dispensing platform. The presented system provides a foundation for future research on perception-assisted medication dispensing, long-term deployment, and clinical validation in smart healthcare environments. Full article

To address the limitations of the centralized hydraulic steering system used in the first-generation heavy-duty wheeled platform developed by our team, this study proposes a fully electrified steering system based on a compact direct-drive electro-mechanical actuator (DEMA) architecture. Compared with the original hydraulic system, the proposed solution reduces the steering-system weight from approximately 150 kg to 32 kg in the single-channel configuration and 40 kg in the dual-channel configuration, while significantly improving system integration and maintainability. For the single-channel DEMA steering system, a composite control strategy combining three-loop PID control with feedforward compensation is developed to improve dynamic response and position-tracking accuracy. AMESim simulation results under a steering resistance torque of 6000 ± 500 Nm show that the system achieves an overshoot below 2%, a steady-state error below 0.1°, and a tracking error below 0.4°. To reduce motor power and thermal-management requirements, a dual-channel DEMA steering architecture is further proposed. Considering inter-channel parameter differences, a primary–secondary synchronization control strategy is developed to suppress force-fighting behavior and improve motion consistency. Simulation results demonstrate that the proposed strategy effectively reduces synchronization errors and maintains highly consistent force output between channels while preserving excellent steering accuracy and tracking performance. The proposed fully electrified steering system and synchronization control strategy provide an effective solution for improving the dynamic performance, lightweight design, and reliability of heavy-duty wheeled platforms. Full article

High-end vertical five-axis machining centers commonly adopt dual-motor direct-drive configurations for their cradle-type A-axis to improve dynamic performance; however, this approach introduces control challenges in balancing counteracting torque and synchronization accuracy due to high-rigidity coupling. To address this issue, this study presents a novel error compensation control strategy based on a synchronous state observer. First, a system dynamic model incorporating dual-axis coupling effects is developed to systematically investigate the coupling mechanism between synchronization error and counteracting torque. Based on this model, a synchronous state observer is designed, which achieves real-time reconstruction and feedforward compensation of synchronization disturbances induced by factors such as transmission parameter mismatches and inter-axis torque imbalance, thereby enabling coordinated control of high-precision position synchronization and torque balance. The effectiveness of the proposed method is verified through simulation and experiments conducted on a VMC630 vertical five-axis machining center. Results show that under various speed and acceleration conditions, the maximum position synchronization error remained below 6.3 × 10^−4∘, with comparable convergence performance; the current deviation between the dual motors was constrained to within ±0.25A, demonstrating effective mitigation of counteracting torque. In machining tests of S-shaped specimens, all measured contour deviations fell within the ±0.060 mm tolerance range, and the specimens exhibited excellent contour consistency and surface quality. These results validate the proposed strategy’s status as an engineering-viable solution for precision motion control in high-rigidity coupled dual-motor systems. Full article

Background/Objectives: Accurate and accessible assessment of upper-limb motion is essential for rehabilitation research, ergonomic evaluation, human–machine interaction, and limb exercise. This work presents a comparative evaluation of upper-limb joint angle estimation obtained from wearable inertial measurement units (IMUs) using a Kinect-based practical benchmark during synchronized data acquisition. Methods: The main variables analyzed were shoulder and elbow joint angles, together with IMU-derived acceleration and surface electromyography (sEMG) signals acquired as complementary physiological information during task execution. Ten healthy adult participants performed predefined upper-limb movements while data from both sensing modalities were recorded simultaneously. Joint angles were estimated independently from IMU and Kinect measurements and compared using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Two One-Sided Tests (TOST) equivalence analysis. Results: For upper- limb motion, IMU-derived estimates showed practical equivalence within the predefined ±10° acceptance margin with small MAE and RMSE values and significant TOST equivalence results (p < 0.001), supporting reliable proximal joint tracking under controlled conditions. Tested elbow motion exhibited large estimation error and large variability, and although the TOST analysis was significant, the equivalence interval slightly exceeded the predefined acceptance bound, indicating comparatively weak agreement between sensing modalities. The presented results should be interpreted as proof-of-concept evidence derived from a comparative benchmark rather than as definitive validation for unrestricted or clinical implementation. The synchronized acceleration and sEMG signals provided complementary temporal information regarding movement execution but were not treated as primary comparative outputs. Conclusions: These findings support the feasibility of wearable IMU-based upper-limb joint angle estimation as a proof-of-concept comparative framework rather than definitive clinical validation. The presented findings support the feasibility of the proposed IMU-based sensing approach for upper-limb joint angle estimation, particularly at the shoulder level, while also highlighting the greater complexity of elbow-related measurements. Further investigation in larger samples, more functionally diverse tasks, and broader populations is required to extend the applicability of the proposed approach. Full article

This study proposes a digital twin-based CPR compression measurement system (DTCMS) architecture for real-time monitoring of CPR compression. The system combines a load cell, an inertial measurement unit (IMU), a LabVIEW acquisition platform, and a CNN module to capture multi-modal motion characteristics during CPR repetitive compression training. A calibration-aware sensor fusion framework synchronizes heterogeneous signals, reduces drift, and enhances robustness under high-frequency operation. Real-time data acquisition, latency-controlled transmission, and digital twin visualization enable synchronized physical–virtual interaction. Experimental results demonstrate high accuracy (R² > 0.99), stable repeatability (coefficient of variation: CV < 3.5%), and reliable dynamic tracking. The compression depth error was maintained within ±1.5 mm, and synchronization latency remained below 0.2 s. Results confirm the proposed DTCMS architecture as a robust solution for real-time biomechanical monitoring and digital twin-based interactive systems. Compared with conventional single-sensor CPR monitoring systems, the proposed framework improves synchronization stability and sensing robustness through calibration-aware multi-sensor fusion. Full article

Submarine landslides pose severe marine geological hazards. Their movement and deposition behaviors can seriously threaten marine engineering stability and coastal safety. The propagation characteristics of landslide-generated tsunamis are therefore critical for hazard assessment. Physical model experiments provide an effective approach for investigating the underlying mechanisms of tsunami generation and propagation. To investigate the complete process from landslide motion to wave generation and propagation, this study developed an underwater soil-movement physical model test system. The system integrates controllable landslide initiation, real-time monitoring of landslide motion, wave height measurements, and full-field image acquisition, enabling synchronous observation of landslide movement and water body response. By controlling the main variables influencing submarine landslide dynamics, a series of physical model experiments were conducted to investigate water surface waves generated under different test conditions. The study examines the complete process from the initial water disturbance caused by submerged landslide motion to tsunami generation and propagation. The effects of landslide volume, particle size, initial submergence depth, and slope angle on tsunami parameters, including wave height, wave velocity, and wave period, were evaluated. Using 21 experimental datasets for each landslide type, namely, cohesionless sandy slides and muddy debris flows, empirical formulas for maximum surge height were established through dimensional analysis, SPSS (v25)-based multiple nonlinear regression, and validation against experimental results. The validation results show strong agreement between the empirical predictions and the physical model test data. Full article

Conventional industrial manipulators are often costly and come with steep learning curves, which limits their scalability in hands-on robotics education. This paper presents a compact and modular vision-guided sorting platform based on a 4-DOF SCARA robot, designed for rapid assembly, reconfiguration, and beginner-friendly deployment in laboratory courses. A collaborative visual perception strategy is proposed, which introduces a lightweight YOLOv8 algorithm for robust material category recognition, while HSV-based color segmentation and Hough circle localization are utilized to extract sub-pixel centroid features. The pixel measurements are mapped to the robot base frame through an integrated nine-point hand–eye calibration model, and joint commands are generated via a joint-space quintic polynomial interpolation algorithm to ensure continuity and avoid kinematic singularities. The overall system adopts a hierarchical architecture in which the vision host communicates target commands to a motion controller via TCP/IP, while joint actuators are driven through a CAN bus. Feasibility is first verified in a Webots digital prototype with synchronized conveyor and manipulator control, and is then validated on a physical platform equipped with a compliant TPU-based soft gripper to improve grasp tolerance under localization noise. Experiments demonstrate that the system achieves an average recognition accuracy of 98.1% and a mean positioning error of 0.189 mm. The proposed platform provides an extensible testbed for teaching kinematics, perception-to-control integration, and modular robotic system development. Full article

The demographic shift towards an aging population necessitates innovative solutions for care and mobility support. While commercial quadruped robots like the Unitree Go1 offer dynamic stability, their native following modes often lack the safety margins and predictability required, and they do not consistently follow the user, at times deviating and navigating independently. This paper presents a robust, vision-based, person-following algorithm designed to address these limitations. Utilizing a ZED 2 stereo camera and Robot Operating System (ROS), the system employs a finite state machine to ensure deterministic target tracking. A velocity control strategy partitions the robot’s motion into distinct stability, proportional, and braking zones based on depth data to ensure fluid interaction. The framework was validated on a Unitree Go1 quadruped platform in an outdoor environment involving 90-degree turns to evaluate tracking robustness. By operating in a headless mode, the system achieved a mean processing latency of $66.5\pm 4.3$ ms. Experimental results demonstrated consistent operational stability, 0.0% intrusion into the intimate safety zone, and effective velocity synchronization between $0.47$ and $0.54$ m/s. While this study establishes a robust technical baseline using healthy subjects, it serves as a preliminary development platform; further iterative testing with elderly users in clinical settings is required to move toward deployment. Beyond the evaluated trials, the framework maintained reliable functional performance across various care facility workshops, successfully following the target in all deployment scenarios. These findings establish a stable technical foundation for the future development of robotic walking partners. Full article

Land subsidence is an important challenge faced by coastal cities under rapid urban development. This study focuses on the Haikou–Laocheng area and conducts time-series monitoring of land subsidence using PS-InSAR and SBAS-InSAR based on 42 Sentinel-1 SAR scenes acquired from April 2023 to April 2025, thereby deriving the spatial distribution of cumulative subsidence rates and the evolution patterns of multi-temporal cumulative subsidence. Because only ascending-orbit Sentinel-1 data were used, the reported deformation values are vertical-projected estimates converted from line-of-sight (LOS) displacement under the assumption that horizontal motion is negligible. The reliability of the monitoring results is evaluated through cross-validation between the two methods, assessing their inter-method consistency. The results indicate that the study area is dominated by slight subsidence, with vertical-projected subsidence rates mainly ranging from −6 to 3.7 mm/y, while a few uplift points are locally observed, forming an overall “stable with localized anomalies” deformation pattern. PS-InSAR and SBAS-InSAR show good consistency in overall trends, and both identify a pronounced subsidence bowl in the southwestern part of the study area, where the peak vertical-projected subsidence rates reach −25.1 mm/y and −35.1 mm/y, respectively, with outward banded attenuation. The results suggest that land subsidence in the study area is influenced by both natural factors and human activities. Specifically, rainfall shows a non-synchronous, stage-wise modulation relationship with subsidence evolution, and most high-subsidence zones are distributed in impervious surfaces such as built-up land and transportation corridors, or in low-elevation areas such as farmland. In terms of geological factors, thick, highly compressible soft soils are the primary geological control on the continued development of subsidence. These findings can provide scientific references for the prevention and control of abnormal subsidence and for urban planning and development in the Haikou–Laocheng area. The strengthened discussion clarifies the research gap, planning significance, and limitations of applying dual time-series InSAR in a data-scarce tropical coastal soft-soil setting. Full article

Digital-avatar systems still provide limited control over emotionally expressive behavior in human–computer interaction, especially in Large Language Model (LLM)-based chatbots and virtual assistants with personalized visual embodiments. To address this problem, we propose Multimodal Avatar Generation (MAVAGEN), a multimodal avatar generation framework for synthesizing upper-body digital avatars with personalized appearance and controllable emotional expression. The user specifies the desired gender and age, as well as provides a short text input from which the target emotional state is inferred. MAVAGEN then retrieves an identity image from the HaGRIDv2-1M corpus and generates an avatar clip with synchronized facial expressions, hand gestures, and expressive speech. The framework uses the following six feature streams: textual features, emotion-distribution features, landmark-based pose features, depth-geometry features, RGB-appearance features, and acoustic features. In a quantitative evaluation against recent human animation methods, MAVAGEN achieves the best overall avatar quality, with FID 48.20, FVD 592.00, SSIM 0.741, Sync-C 7.40, HKC 0.929, HKV 25.30, CSIM 0.563, and EmoAcc 0.88. Ablation results show that emotion and acoustic features contribute most to emotional agreement, while landmark-based pose and depth features improve geometric and motion stability. These results support the practical use of MAVAGEN in personalized LLM-based assistants and other emotion-sensitive interactive systems. Full article

This study aims to enhance the precision of humanoid robots in imitating complex human “walking–grasping” coordinated movements. Addressing limitations in sample efficiency and reward function design in Generative Adversarial Imitation Learning (GAIL), we propose the Similarity Reward-Augmented Generative Adversarial Imitation Learning (SRA-GAIL) framework. The method integrates plantar thin-film resistive pressure sensors to measure the real-time pressure distribution at four key points on both feet, combined with roll/pitch angle data acquired from JY901S inertial measurement units (IMUs). A Lagrangian constraint optimization strategy is employed to achieve gait stability control based on the zero moment point (ZMP). Simultaneously, a visual similarity evaluation module is established using human demonstration trajectories captured by a Logitech C920E camera, augmented by grip force feedback from flexible thin-film pressure sensors on the hands. This enables the design of a multimodal sensor-fused similarity reward function. By incorporating Lagrangian constraint optimization and a maximum entropy reinforcement learning framework, Similarity Reward-Augmented Generative Adversarial Imitation Learning synchronously optimizes gait stability control—guided by zero moment point (ZMP) and roll/pitch data—and vision-based trajectory similarity evaluation. These components address motion stability constraints and trajectory similarity metrics, respectively, generating biomechanically plausible gait strategies. A spatiotemporal attention mechanism parses human motion trajectory features to drive the end-effector for high-precision trajectory tracking. To validate the proposed method, an imitation learning experimental system was constructed on a physical XIAOLI humanoid robot platform, integrating inertial measurement units (IMUs), plantar pressure sensors, and a vision system. Quantitative evaluations were conducted across multiple dimensions, including robot platform analysis, walking stability, object grasping success rates, and end-effector trajectory similarity. The results demonstrate that, compared to Generative Adversarial Imitation Learning (GAIL) and behavioral cloning, Similarity Reward-Augmented Generative Adversarial Imitation Learning achieves a stable object grasping success rate of 93.7% in complex environments, with a 23.8% improvement in sample efficiency. The method maintains a 96.5% compliance rate for zero moment point (ZMP) trajectories within the support polygon, significantly outperforming baseline approaches. This effectively addresses the bottleneck in robot policies adapting to dynamic changes in real-world environments. Full article

Hybrid CNC and additive manufacturing platforms often rely on host-assisted or otherwise overdimensioned control architectures to achieve deterministic multi-axis motion, increasing system cost and complexity. This paper presents a fully microcontroller-based timer–DMA motion execution architecture that eliminates the need for external processors or FPGA-based execution, enabling deterministic multi-axis synchronization under the tested conditions in a simpler, more cost-effective way. The proposed framework integrates motion planning, precise step-time computation, and hardware-assisted pulse generation within a unified embedded control architecture. The main novelty lies in the systematic use of timer and DMA peripherals to offload time-critical pulse execution from the microcontroller core, allowing it to focus on motion planning and precise step-time computation. Unlike segmentation-based approaches, the duration of each individual step is calculated directly without fixed-interval segmentation, enabling high motion resolution while avoiding per-step interrupts that introduce jitter at high motion speeds. The architecture was validated on a hybrid platform capable of both milling and material extrusion. Experimental results confirmed real-time feasibility within practical on-chip memory limits and demonstrated very small interpolation errors caused mainly by timer quantization, comparable to those observed in host-processor-based motion systems. Machining and additive-manufacturing experiments further confirmed stable execution and accurate trajectory tracking under real operating conditions. Full article

When a rigid object is manipulated by dual arms to form a closed chain, the dual-arm motion must satisfy closed-chain constraints. Although synchronized motion can be achieved by strictly tracking predefined global trajectories, the presence of dynamic obstacles necessitates reactive local planning. However, existing local planning methods designed for single-arm manipulators cannot guarantee synchronization between dual arms. To address this limitation, we propose a dual-arm reactive synchronized motion controller (SMC) by incorporating closed-chain constraints on dual-arm slack velocities based on spherical geometric velocity constraints, and by implementing a flexible master-slave arm switching strategy. As a result, the proposed controller achieves synchronized dual-arm control while preserving excellent motion performance, including manipulability enhancement, obstacle avoidance, and compliance with joint angle and velocity constraints. Simulations and experiments on a humanoid upper-body robot validate the effectiveness of the proposed approach. Full article

The torque ripples of robotic permanent magnet synchronous motors (PMSMs) degrade motion smoothness and positioning accuracy of the system, while inevitable load transients in robotic tasks further complicate torque ripple attenuation. To address this issue, this paper develops an event-triggered torque ripple attenuation method that explicitly distinguishes torque ripple from dynamic load transients. First, a sliding-mode torque observer is constructed to obtain real-time torque information, whose stability is rigorously analyzed using a Lyapunov function. Second, frequency-selective torque ripple extraction schemes are proposed to accurately isolate steady-state high-frequency torque ripple from the estimated torque signal. In particular, two specially designed filtering structures are developed and compared, one of which is selected to preserve ripple-related frequency content during test, ensuring robust and accurate ripple identification under varying operating conditions in robotics. Third, a torque-ripple-regulation-based compensation strategy is used within a vector-controlled PMSM drive, in which the extracted torque ripple is processed by a dedicated ripple regulator to generate voltage compensation signals. This strategy achieves effective steady-state torque ripple attenuation with low implementation complexity, while avoiding performance degradation during dynamic load transients. Finally, experimental results are provided to validate the effectiveness of the proposed methods. Full article

Anterior cruciate ligament (ACL) injury risk is elevated in female youth football, yet knee joint loading has mainly been studied under controlled laboratory conditions. This limits understanding of how injury risk emerges during realistic match situations. This study provided a field-based kinetic characterization of football-specific movements by estimating knee abduction moments (KAMs) using wearable sensors and machine learning. Fifty-two highly talented female youth players performed agility tasks during training, including structured exercises (F-EX) and game-based play (F-GAME). Full-body kinematics were collected with inertial measurement units, and a validated support vector machine model, trained on synchronized motion capture and force plate data, classified trials as high or low KAM. Across 662 change-in-direction trials, 9–12% were classified as high KAM in both conditions, indicating that potentially high-risk loading regularly occurs during routine actions. High KAM trials showed reduced knee and pelvis flexion, increased hip flexion, and greater pelvis rotation toward the cutting direction, reflecting upright, stiff movement strategies. Performance analyses revealed smaller cut angles in exercises and greater approach acceleration in game play, without differences in peak velocity. These findings demonstrate the feasibility of field-based kinetic screening and support a complex-systems perspective on ACL injury risk. Full article