Search Results (1,760)

Cattle behavior is closely related to their health status, and monitoring cattle behavior using intelligent devices can assist herders in achieving precise and scientific livestock management. Current behavior recognition algorithms are typically executed on server platforms, resulting in increased power consumption due to data transmission from edge devices and hindering real-time computation. An edge-based cattle behavior recognition method via Graph Neural Network (GNN) compression is proposed in this paper. Firstly, this paper proposes a wearable device that integrates data acquisition and model inference. This device achieves low-power edge inference function through a high-performance embedded microcontroller. Secondly, a sequential residual model tailored for single-frame data based on Inertial Measurement Unit (IMU) and displacement information is proposed. The model incrementally extracts deep features through two Residual Blocks (Resblocks), enabling effective cattle behavior classification. Finally, a compression method based on GNNs is introduced to adapt edge devices’ limited storage and computational resources. The method adopts GNNs as the backbone of the Actor–Critic model to autonomously search for an optimal pruning strategy under Floating-Point Operations (FLOPs) constraints. The experimental results demonstrate the effectiveness of the proposed method in cattle behavior classification. Moreover, enabling real-time inference on edge devices significantly reduces computational latency and power consumption, thereby highlighting the proposed method’s advantages for low-power, long-term operation. Full article

Step detection based on measurements of inertial measurement units (IMUs) is fundamental for human activity recognition, indoor navigation, and health monitoring applications. This study validates and compares five fundamentally different step detection algorithms for potential implementation on edge devices. A dedicated measurement system based on the Raspberry Pi Pico 2W microcontroller with two IMU sensors (Waveshare Pico-10DOF-IMU and Adafruit ST-9-DOF-Combo) was designed. The implemented algorithms include Peak Detection, Zero-Crossing, Spectral Analysis, Adaptive Threshold, and SHOE (Step Heading Offset Estimator). Validation was performed across 84 measurement sessions covering seven test scenarios (Timed Up and Go test, natural and fast walking, jogging, and stair climbing) and four sensor mounting locations (thigh pocket, ankle, wrist, and upper arm). Results demonstrate that Peak Detection achieved the best overall performance, with an average F1-score of 0.82, while Spectral Analysis excelled in stair scenarios (F1 = 0.86–0.92). Surprisingly, upper arm mounting yielded the highest accuracy (F1 = 0.84), outperforming ankle placement. The TUG clinical test proved most challenging (average F1 = 0.68), while fast walking was easiest (F1 = 0.87). Additionally, a preliminary application to 668 clinical TUG recordings from the open-access FRAILPOL database revealed algorithm-specific failure modes when continuous gait assumptions are violated. These findings provide practical guidelines for algorithm selection in edge computing applications and activity monitoring systems. Full article

Agitation-related movements in intensive care units (ICUs), such as unintended tube removal and bed exit attempts, pose significant risks to patient safety. The wearable inertial measurement units (IMUs) offer a potential means of capturing such movements. However, the technical feasibility of discriminating ICU-relevant agitation motion patterns from multi-site IMU data remains insufficiently established. To evaluate the technical feasibility of using a convolutional neural network (CNN) applied to multi-site IMU signals to discriminate predefined ICU-relevant agitation-related motion patterns under controlled experimental conditions. Fifteen healthy volunteers performed six scripted movements designed to emulate ICU-relevant agitation-related behaviors while wearing seven IMU sensors on the limbs and waist. A CNN comprising three convolutional layers with kernel sizes of 3, 5, and 7 was trained using 1-s windows extracted from 8-s trials and evaluated using leave-one-subject-out cross-validation. The performance was summarized using macro-averaged accuracy, sensitivity, specificity, precision, and F1 score. Across 135 independent training runs, the CNN achieved a median macro-averaged accuracy of 77.0%, sensitivity of 77.0%, specificity of 95.4%, and F1 score of 77.4%. These results indicate stable window-level discrimination of the predefined motion classes under standardized conditions. This proof-of-concept study demonstrates that multi-site IMU signals combined with CNN-based modeling can technically discriminate ICU-relevant agitation-related motion patterns in a controlled laboratory setting. Although these findings do not establish clinical validity in ICU patients, they provide a methodological foundation for future studies aimed at patient-level validation and real-world critical care deployment. Full article

Falls are a prominent issue in society and the second leading cause of unintentional death globally. Traditional gait analysis is a process that can aid in identifying factors that increase a person’s risk of falling through determining their gait parameters in a controlled environment. Advances in wearable sensor technology and analytical methods such as deep learning can enable remote gait analysis, increasing the quality of the collected data, standardizing the process between centers, and automating aspects of the analysis. Real-world gait analysis requires two problems to be solved: high-accuracy Human Activity Recognition (HAR) and high-accuracy terrain classification. High accuracy HAR has been achieved through the application of powerful novel classification techniques to various HAR datasets; however, terrain classification cannot be approached in this way due to a lack of suitable datasets. In this study, we present the Context-Aware Human Activity Recognition (CAHAR) dataset: the first activity- and terrain-labeled dataset that targets a full range of indoor and outdoor terrains, along with the common gait activities associated with them. Data were captured using Inertial Measurement Units (IMUs), Force-Sensing Resistor (FSR) insoles, color sensors, and LiDARs from 20 healthy participants. With this dataset, researchers can develop new classification models that are capable of both HAR and terrain identification to progress the capabilities of wearable sensors towards remote gait analysis. Full article

Female soccer is rapidly becoming a widely practiced sport at different levels: this opens up a new demand for systems meant to protect athletes from head impacts or to monitor their effects. The market is offering some solutions in similar sports, but the specificity and high relevance of soccer encourage the development of a dedicated solution. From market analysis, technology scouting, and ethnographic research a set of functional and technical requirements have been defined and proposed. The designed instrumented head band is equipped with one Inertial Measurement Unit (IMU) in the occipital area and four contact pressure sensors on the sides. The concept design is low-cost and open-architecture, prioritizing accessibility over complexity. The modularity also ensures that each component (sensing, battery, communication) can be replaced or upgraded independently, enabling iterative refinement and integration into future sports safety systems. In addition to safety monitoring for injury prevention or detection of the traumatic impact, the system is relevant for supporting performance monitoring, rehabilitation or post-injury recovery and other important applications. System engineering has started and the next step is building the prototypes for testing and validation. Full article

Hip osteoarthritis (OA) affects the entire joint and significantly alters gait. Assessing gait through a single trunk inertial measurement unit (IMU) in clinical settings offers a more practical alternative to complex laboratory settings, allowing for the capture of natural gait movements with valuable biomechanical insights. We evaluated (1) whether gait quality differs between individuals with hip OA and healthy controls during habitual and fast walking, (2) whether gait changes from habitual to fast walking differ between groups. Forty individuals with hip OA and 40 age-matched healthy controls underwent 25-m habitual walk and 40-m fast walk. Six gait quality parameters—step symmetry, stride symmetry, stability, smoothness, regularity, and complexity—were analyzed from the IMU signals. During habitual walking, individuals with hip OA exhibited reduced symmetry and stability and several vertical impairments. During fast walking, individuals with hip OA continued to show reduced step symmetry and a more constrained gait in the mediolateral direction. Additionally, people with hip OA also showed limited adjustments when transitioning from habitual to fast walking, in contrast to the significant adjustments observed in healthy controls. These findings indicate that gait in individuals with hip OA is impaired during habitual and fast walking, with limited adaptations across the transition between the two conditions. Full article

Recent reports have revealed that the number of lower limb amputees worldwide has increased as a result of war, accidents, and vascular diseases and that transfemoral amputation accounts for 39% of cases, highlighting the need to develop an improved functional prosthetic knee joint that improves the amputee’s ability to resume activities of daily living. To enable transfemoral prosthesis users to walk on level ground, accurate prediction of the intended knee joint angle is critical for transfemoral prosthesis control. Therefore, the purpose of this research was to develop a technique for estimating knee joint angle utilizing a long short-term memory (LSTM) network and kinematic data collected from inertial measurement units (IMUs). The proposed LSTM network was trained and tested to estimate the contralateral knee angle using data collected from twenty able-bodied subjects using a lab-developed sensory gadget, which included four IMUs. Accordingly, the present work represents a feasibility investigation conducted on able-bodied individuals rather than a clinical validation for amputee gait. This study contributes to the field of bionics by mimicking the natural biomechanical behavior of the human knee joint during gait cycle to improve the control of artificial prosthetic knees. The proposed LSTM model learns the contralateral knee’s motion patterns in able-bodied gait and demonstrates the potential for future application in prosthesis control, although direct generalization to amputee users is outside the scope of this preliminary study. The contralateral LSTM models exhibited a real-time RMSE range of 2.48–2.78° and a correlation coefficient range of 0.9937–0.9991. This study proves the effectiveness of LSTM networks in estimating contralateral knee joint angles and shows their real-time performance and robustness, supporting its feasibility while acknowledging that further testing with amputee participants is required. Full article

This paper presents a virtual reference-based preview semi-active suspension system using a Magneto-Rheological (MR) damper to improve ride comfort when traversing bumps. The algorithm is designed to track the virtual reference profile of the vehicle’s corner by introducing a Model Predictive Control (MPC) method while considering the passivity of the MR damper. The proposed MPC is formulated to rely solely on estimable variables from an Inertial Measurement Unit (IMU) and vertical accelerometer. To support implementation on an Electronic Control Unit (ECU), the suspension state estimator employs a simple band-limited filtering structure. The proposed method is evaluated in simulation and achieves performance comparable to a controller that has accurate prior knowledge of the road profile. In addition, simulation results demonstrate that the proposed approach exhibits low sensitivity to sensor noise and bump perception uncertainty, making it well suited for real-world vehicle applications. Full article

To address the urgent need for high-fidelity motion data for validating navigation algorithms for Unmanned Underwater Vehicles (UUVs), this paper proposes a data generation method based on a parametric motion model. First, based on the principles of rigid body dynamics and fluid mechanics, a decoupled six-degrees-of-freedom (6-DOF) Linear and Angular Acceleration Vector (LAAV) model is constructed, establishing a dynamic mapping relationship between the rudder angle and speed setting commands and motion acceleration. Second, a segmentation–identification framework is proposed for three-dimensional trajectory segmentation, integrating Gaussian Process Regression and Ordering Points To Identify the Clustering Structure (GPR-OPTICS), along with a Dynamic Immune Genetic Algorithm (DIGA). This framework utilizes real vessel data to achieve motion segment clustering and parameter identification, completing the construction of the LAAV model. On this basis, by introducing sensor error models, highly credible Inertial Measurement Unit (IMU) data are generated, and a complete attitude, velocity, and position (AVP) motion sequence is obtained through an inertial navigation solution. Experiments demonstrate that the AVP data generated by our method achieve over 88% reliability compared with the real vessel dataset. Furthermore, the proposed method outperforms the PSINS toolbox in both the reliability and accuracy of all motion parameters. These results validate the effectiveness and superiority of our proposed method, which provides a high-fidelity data benchmark for research on underwater navigation algorithms. Full article

This study presents a standardized method for multimodal monitoring of exercise execution in isolated, confined, and extreme (ICE) environments, addressing the need for reproducible assessment of neuromuscular and cardiovascular responses under space- and equipment-limited conditions. The method integrates wearable surface electromyography (sEMG), inertial measurement units (IMU), and electrocardiography (ECG) to capture muscle activation, movement, and cardiac dynamics during space-efficient exercise. Ten exercises suitable for confined habitats were implemented during analog missions conducted in the DeepLabH03 facility, with feasibility evaluated in a seven-day campaign involving three adult participants. Signals were synchronized using video-verified repetition boundaries, sEMG was normalized to maximum voluntary contraction, and sEMG amplitude- and frequency-domain features were extracted alongside heart rate variability indices. The protocol enabled stable real-time data acquisition, reliable repetition-level segmentation, and consistent detection of muscle-specific activation patterns across exercises. While amplitude-based sEMG indices showed no uniform main effect of exercise, robust exercise-by-muscle interactions were observed, and sEMG mean frequency demonstrated sensitivity to differences in movement strategy. Cardiac measures showed limited condition-specific modulation, consistent with short exercise bouts and small sample size. As a proof-of-concept feasibility study, the proposed protocol provides a practical and reproducible framework for multimodal physiological monitoring of exercise in ICE analogs and other constrained environments, supporting future studies on exercise quality, training load, and adaptive feedback systems. The protocol is designed to support near-real-time monitoring and forms a technical basis for future exercise-quality feedback in confined habitats. Full article

Aiming to address the safety and stability issues caused by typical faults of Unmanned Aerial Vehicle (UAV) navigation sensors, a novel fault-tolerant method is proposed, which can capture the temporal dependencies of fault feature evolution, and complete the classification, prediction, and data reconstruction of fault data. In this fault-tolerant method, the feature extraction module adopts the FNKPCA method—integrating the Frobenius Norm (F-norm) with Kernel Principal Component Analysis (KPCA)—to optimize the kernel function’s ability to capture signal features, and enhance the system reliability. By combining FNKPCA with Support Vector Machine (SVM) and Bidirectional Long Short-Term Memory (BiLSTM), an active fault-tolerant processing method, namely FNKPCA–SVM–BiLSTM, is obtained. This study conducts comparative experiments on public datasets, and verifies the effectiveness of the proposed method under different fault states. The proposed approach has the following advantages: (1) It achieves a detection accuracy of 98.64% for sensor faults, with an average false alarm rate of only 0.15% and an average missed detection rate of 1.16%, demonstrating excellent detection performance. (2) Compared with the Long Short-Term Memory (LSTM)-based method, the proposed fault-tolerant method can reduce the RMSE metrics of Global Positioning System (GPS), Inertial Measurement Unit (IMU), and Ultra-Wide-Band (UWB) sensors by 77.80%, 14.30%, and 75.00%, respectively, exhibiting a significant fault-tolerant effect. Full article

In the realm of wearable technology, achieving robust continuous authentication requires balancing high security with the strict resource constraints of embedded platforms. Conventional machine learning approaches and deep learning-based biometrics often incur high computational costs, making them unsuitable for low-power edge devices. To address this challenge, we propose H-PPG, a lightweight authentication system that integrates photoplethysmography (PPG) and inertial measurement unit (IMU) signals for continuous user verification. Using Hyperdimensional Computing (HDC), a lightweight classification framework inspired by brain-like computing, H-PPG encodes user physiological and motion data into high-dimensional hypervectors that comprehensively represent individual identity, enabling robust, efficient and lightweight authentication. An adaptive learning process is employed to iteratively refine the user’s hypervector, allowing it to progressively capture discriminative information from physiological and behavioral samples. To further enhance identity representation, a dimension regeneration mechanism is introduced to maximize the information capacity of each dimension within the hypervector, ensuring that authentication accuracy is maintained under lightweight conditions. In addition, a user-defined security level scheme and an adaptive update strategy are proposed to ensure sustained authentication performance over prolonged usage. A wrist-worn prototype was developed to evaluate the effectiveness of the proposed approach and extensive experiments involving 15 participants were conducted under real-world conditions. The experimental results demonstrate that H-PPG achieves an average authentication accuracy of 93.5%. Compared to existing methods, H-PPG offers a lightweight and hardware-efficient solution suitable for resource-constrained wearable devices, highlighting its strong potential for integration into future smart wearable ecosystems. Full article

The use of ships and boats as sea-state (SS) measurement platforms has the potential to expand ocean observations while providing actionable information for real-time operational decision-making at sea. Within the framework of the Wave Buoy Analogy (WBA), this work develops an inverse approach in which efficient simulations of wave-induced motions of an advancing vessel are used to train a neural network (NN) to predict SS parameters across a broad range of wave climates. We show that a reduced set of novel motion discriminant variables (MDVs)—computed from short time series of heave, roll, and pitch motions measured by an onboard inertial measurement unit (IMU), together with the vessel’s forward speed—provides sufficient and robust information for accurate, near-real-time SS estimation. The methodology targets small, barge-like tugboats whose operations are SS-limited and whose motions can become large and strongly nonlinear near their upper operating limits. To accurately model such responses and generate training data, an efficient nonlinear time-domain seakeeping model is developed that includes nonlinear hydrostatic and viscous damping terms and explicitly accounts for forward-speed effects. The model is experimentally validated using a scaled physical model in laboratory wave-tank tests, demonstrating the necessity of these nonlinear contributions for this class of vessels. The validated model is then used to generate large, high-fidelity datasets for NN training. When applied to independent numerically simulated motion time series, the trained NN predicts SS parameters with errors typically below 5%, with slightly larger errors for SS directionality under relatively high measurement noise. Application to experimentally measured vessel motions yields similarly small errors, confirming the robustness and practical applicability of the proposed framework. In operational settings, the trained NN can be deployed onboard a tugboat and driven by IMU measurements to provide real-time SS estimates. While results are presented for a specific vessel, the methodology is general and readily transferable to other ship geometries given appropriate hydrodynamic coefficients. Full article

Background: We explored a new approach for increasing participation in daily life for individuals with severe movement impairments. The core of the approach is an application designed to Translate Intentions into Action (TIA) as a motivational tool for both leisure and clinical training sessions. Methods: The TIA application was developed to enable users to activate motivational feedback, like sounds, music, or videos, through movement measured with an IMU (Inertial Measurement Unit). IMUs were calibrated to user-specific thresholds based on individual movement potential. TIA was tested in two different age groups to assess applicability throughout lifespan and across different motor capacities. Results: The results indicated that TIA can be used for improved participation when positive feedback is provided during the intervention sessions. Observations demonstrated that regardless of age and motor capabilities, increased participation was achieved. TIA demonstrated the far-reaching potential to enhance the engagement and motivation of individuals with different levels of severe disabilities. Conclusions: By providing personalized, positive feedback through movement-activated outputs, TIA can be used by a wide range of people, with or without motor disabilities, to control digital outputs, such as video and audio. These findings suggest that TIA can be a valuable tool in both clinical and leisure settings to promote meaningful participation in activities. Full article

Wearable inertial measurement units (IMUs) provide an accessible means of monitoring fatigue-related changes in running biomechanics, yet most existing methods rely on limited feature sets, lack personalization, or fail to generalize across individuals. This study introduces a bioinformatics-inspired stride sequence modeling framework that integrates spectral–entropy features, sample entropy, frequency-domain descriptors, and mixed-effects statistical modeling to detect fatigue using a single lumbar-mounted IMU. Nineteen recreational runners completed non-fatigued and fatigued 400 m runs, from which we extracted stride-level features and evaluated (1) population-level fatigue classification via global leave-one-participant-out (LOPO) models and (2) individualized fatigue detection through supervised participant-specific models and non-fatigued-only anomaly detection. Mixed-effects models revealed robust and multidimensional fatigue effects across key biomechanical features, with large standardized effect sizes (Cohen’s d up to 1.35) and substantial variance uniquely explained by fatigue (partial R² up to 0.31). Global LOPO machine learning models achieved modest accuracy (55%), highlighting strong inter-individual variability. In contrast, personalized supervised Random Forest classifiers achieved near-perfect performance (mean accuracy 97.7%; mean AUC 0.997), and NF-only One-Class SVMs detected fatigue as a deviation from individual baseline patterns (mean AUC 0.967). Entropy and stride-to-stride variability metrics further demonstrated consistent fatigue-linked increases in movement irregularity and reduced neuromuscular control. These findings show that IMU stride sequences contain highly informative, fatigue-sensitive biomechanical signatures, and that combining bioinformatics-inspired sequence analysis with hybrid statistical and personalized AI models enables both robust population-level insights and highly reliable individualized fatigue monitoring. The proposed framework supports future integration into sports analytics platforms, digital coaching systems, and real-time wearable fatigue detection technologies. This highlights the necessity of personalized fatigue-monitoring strategies in wearable systems. Full article