Search Results (281)

Motion capture (MoCap) data derived from wearable Inertial Measurement Units is essential to applications in sports science and healthcare robotics. However, a significant amount of the potential of this data is limited due to missing data derived from sensor limitations, network issues, and environmental interference. Such limitations can introduce bias, prevent the fusion of critical data streams, and ultimately compromise the integrity of human activity analysis. Despite the plethora of data imputation techniques available, there have been few systematic performance evaluations of these techniques explicitly for the time series data of IMU-derived MoCap data. We address this by evaluating the imputation performance across three distinct contexts: univariate time series, multivariate across players, and multivariate across kinematic angles. To address this limitation, we propose a systematic comparative analysis of imputation techniques, including statistical, machine learning, and deep learning techniques, in this paper. We also introduce the first publicly available MoCap dataset specifically for the purpose of benchmarking missing value imputation, with three missingness mechanisms: missing completely at random, block missingness, and a simulated value-dependent missingness pattern simulated at the signal transition points. Using data from 53 karate practitioners performing standardized movements, we artificially generated missing values to create controlled experimental conditions. We performed experiments across the 53 subjects with 39 kinematic variables, which showed that discriminating between univariate and multivariate imputation frameworks demonstrates that multivariate imputation frameworks surpassunivariate approaches when working with more complex missingness mechanisms. Specifically, multivariate approaches achieved up to a 50% error reduction (with the MAE improving from 10.8 ± 6.9 to 5.8 ± 5.5) compared to univariate methods for transition point missingness. Specialized time series deep learning models (i.e., SAITS, BRITS, GRU-D) demonstrated a superior performance with MAE values consistently below 8.0 for univariate contexts and below 3.2 for multivariate contexts across all missing data percentages, significantly surpassing traditional machine learning and statistical methods. Notable traditional methods such as Generative Adversarial Imputation Networks and Iterative Imputers exhibited a competitive performance but remained less stable than the specialized temporal models. This work offers an important baseline for future studies, in addition to recommendations for researchers looking to increase the accuracy and robustness of MoCap data analysis, as well as integrity and trustworthiness. Full article

Inertial Measurement Units (IMUs) offer promising alternatives to traditional motion capture systems, especially in real-world sports scenarios. Accurate foot contact detection (FCD) is crucial for biomechanical analysis, and since on-the-field force plates are unsuitable, IMU-based FCD algorithms have been extensively investigated. However, sports activities leading to musculoskeletal injuries are multidirectional and high-dynamics in nature and FCD algorithms, which have mostly been studied in gait analysis, might sensibly worsen performance. This systematic review (PROSPERO, ID: CRD420251010584) aimed to evaluate IMU-based FCD algorithms applied to high-dynamics sports tasks, identifying strengths, limitations, and areas for improvement. A multi-database search was conducted until May 2025. Studies were included if they applied IMU-based FCD algorithms in high-dynamic movements. In total, 37 studies evaluating 71 FCD algorithms were included. Most papers focused on running, with only 3 on cut manoeuvres. Almost all studies involved healthy individuals only, and foot linear acceleration was the most inspected FCD metric. FCD algorithms demonstrated high accuracy, though speed variation impacted performance in 23/37 studies. This review highlights the lack of validated IMU-based FCD algorithms for high-dynamic sports movements and emphasizes the need for improved methods to advance sports biomechanics testing in injury prevention. Full article

This paper presents the development of an optoelectronic wearable sensor system for portable monitoring of the movement and physiological parameters of patients. The sensor system is based on a low-cost inertial measurement unit (IMU) and an optical fiber-integrated chest belt for breathing rate monitoring with wireless connection with a gateway connected to the cloud. The sensors also use artificial intelligence algorithms for clustering, classification, and regression of the data. Results show a root mean squared error (RMSE) between the reference data and the proposed breathing rate sensor of 0.6 BPM, whereas RMSEs of 0.037 m/s² and 0.27 °/s are obtained for the acceleration and angular velocity analysis, respectively. For the sensor validation under different movement analysis protocols, the balance and Timed up and Go (TUG) tests performed with 12 subjects demonstrate the feasibility of the proposed device for biomechanical and physical therapy protocols’ automatization and assessment. The balance tests were performed in two different conditions, with a wider and narrower base, whereas the TUG tests were made with the combination of cognitive and motor tests. The results demonstrate the influence of the change of base on the balance analysis as well as the dual task effect on the scores during the TUG testing, where the combination between motor and cognitive tests lead to smaller scores on the TUG tests due to the increase of complexity of the task. Therefore, the proposed approach results in a low-cost and fully automated sensor system that can be used in different protocols for physical rehabilitation. Full article

As advancements in inertial measurement units (IMUs) for motion analysis progress, the inability to directly apply decades of research-based optical motion capture (OMC) methodologies presents a significant challenge. This study aims to bridge this gap by proposing an innovative deep learning approach to predict marker positions from IMU data, allowing traditional OMC-based calculations to estimate joint kinematics. Eighteen participants walked on a treadmill with seven IMUs and retroreflective markers. Trials were divided into normalized gait cycles (101 frames), and an autoencoder network with a custom Biomech loss function was used to predict 16 marker positions from IMU data. The model was validated using the leave-one-subject-out method and assessed using root mean squared error (RMSE). Joint angles in the sagittal plane were calculated using OMC methods, and RMSE was computed with and without alignment using dynamic time warping (DTW). The models were also tested on external datasets. Marker predictions achieved RMSE values of 2–4 cm, enabling joint angle predictions with 4–7° RMSE without alignment and 2–4° RMSE after DTW for sagittal plane joint angles (ankle, knee, hip). Validation using separate and open-source datasets confirmed the model’s generalizability, with similar RMSE values across datasets (4–7° RMSE without DTW and 2–4° with DTW). This study demonstrates the feasibility of applying conventional biomechanical models to IMUs, enabling accurate movement analysis and visualization outside controlled environments. This approach to predicting marker positions helps to bridge the gap between IMUs and OMC systems, enabling decades of research-based biomechanical methodologies to be applied to IMU data. Full article

The Wolf Motor Function Test (WMFT) is widely used to evaluate upper limb motor performance after stroke. However, conventional approaches may overlook domain-specific neuromuscular and kinematic differences during task execution. This study classified WMFT tasks into three functional domains: proximal reaching and transport (PRT), fine motor manipulation (FMM), and gross motor functional control (GMFC). Interlimb differences in muscle activation, joint mobility, and movement amplitude were examined using sensor-based measurements. Twelve individuals with chronic stroke performed 16 WMFT tasks. Surface electromyography (EMG) and inertial measurement units (IMUs) recorded upper limb muscle activity, joint angles, and segmental displacement. Wilcoxon signed-rank tests and Spearman correlations were conducted for each functional domain. Significant asymmetries in EMG, range of motion (ROM), and root mean square (RMS) acceleration were found in PRT and FMM tasks. These results reflect increased proximal muscle activation and reduced distal engagement on the paretic side. GMFC tasks elicited more symmetrical patterns but still showed subtle deficits in distal control. Correlation analyses demonstrated strong interdependencies among neuromuscular and kinematic measures. This finding underscores the integrated nature of compensatory strategies. Categorizing WMFT tasks by functional domain and integrating multimodal sensor analysis revealed nuanced impairment patterns. These patterns were not detectable by conventional observational scoring. These findings support the use of sensor-based, domain-specific assessment to guide individualized rehabilitation strategies. Such approaches may ultimately enhance long-term functional recovery in stroke survivors. Full article

The purpose of the present study was twofold: (i) to examine within-group differences in external load metrics during practice and official competition, and (ii) to examine between-group differences in external load metrics across the U16 and U18 levels of play. A total of thirty-six female athletes participated in the present study, of which nineteen were U16 and seventeen were U18 basketball players. The athletes wore an inertial measurement unit system (Kinexon) sampling at 20 Hz during practice and official games. The average values for each external load metric across ten practices and five games were used for performance analysis. Dependent and independent t-tests were used to examine within- and between-group statistically significant differences, respectively (p < 0.05). The findings reveal that the external load placed on the athletes during the game (e.g., distance covered, average speed, total number of accelerations and decelerations) was considerably greater than the external load during practice sessions, both on the U16 and U18 levels of play. Conversely, while the game-induced external load remained consistent across the two competitive levels, U18 players tended to spend more time and cover more distance in low-speed zones than in high-speed zones during practice, compared to their U16 counterparts, suggesting their superior movement efficiency. Full article

The assessment of physical quantity values, especially in case of sports-related activities, is critical to evaluate the performance and fitness level of athletes. In real-world applications, motion analysis tools are often employed to assess motor performance in subjects. In case the methods used to calculate a specific quantity of interest differ from each other, different values may be provided as output. Therefore, there is the need to get a coherent final measurement, giving the possibility to compare results homogeneously, combining the different methodologies used by the instruments. These tools vary in measurement capabilities and the physical principles underlying the measurement procedures. Emerging differences in results could lead to non-uniform evaluation metrics, thus making a fair comparison unpracticable. A possible solution to this problem is provided in this paper by implementing an iterative approach, working on two measurement time series acquired by two different instruments, specifically focused on jump height estimation. In the analyzed case study, two instruments estimate the jump height exploiting two different technologies: the inertial and the vision-based ones. In the first case, the measurement value depends on the movement of the center of gravity during jump activity, while, in the second case, the jump height is derived by estimating the maximum distance ground–foot during the jump action. These approaches clearly could lead to different values, also considering the same jump test, due to their observation point. The developed methodology can provide three different ways out: (i) mapping the inertial values towards the vision-based reference system; (ii) mapping the vision-based values towards the inertial reference system; (iii) determining a comprehensive measurement, incorporating both contributions, thus making measurements comparable in time (performance progression) and space (comparison among subjects), eventually adopting only one of the analyzed instruments and applying the transformation algorithm to get the final measurement value. Full article

Objectives. The present study aimed to quantify changes in spinal kinematics during Nordic walking compared to regular walking (RW) for 60 s on a training path among physically fit young males (n = 20, aged 19–22 years). Methods. Two walking speeds were analyzed: 3 km/h and 5 km/h. The experimental setup was designed to assess spinal angular rotations using five kinematic parameters: upper spine, lower spine, thoracic region, lumbar region, and pelvis. Results. The data were acquired from 9 compact inertial sensors and the following motion analysis is carried out using 3D MioMotion IMU sensor’s analysis system. The differences in the obtained cyclic biomechanical parameters were detected using functional data analysis (FDA) statistical tests. Conclusions. The key finding of the study is that Nordic walking significantly alters the angular kinematic pattern of spinal movement as it revealed significant differences in all five measured parameters when compared to normal walking. Notably, the most pronounced changes were observed in the upper spine and pelvis motion. Additionally, Nordic walking increased stance phase duration and velocity: (i) significantly increased the duration of the stance phase in all three planes of motion; (ii) significantly increased the velocity during the stance phase across all three planes. These reported findings highlight the biomechanical, preventive, therapeutic, and rehabilitative potential of Nordic walking. Full article

Background: This study examined the influence of creative thinking, shaped by different forms of episodic mental representations, on human movement. The primary objective was to investigate how creativity, elicited through distinct cognitive stimuli, affects movement variability. Methods: Twenty-four professional dancers developed two original dance phrases, each inspired by either a visual or a narrative mental representation. Movement data were collected via inertial sensor technology and subsequently analysed to determine differences in motor expression. Results: The results indicated that movements performed under narrative representation conditions exhibited significantly increased risk-taking behaviour, greater movement amplitude, and a higher overall movement volume compared to those guided by visual stimuli. Conclusions: These findings underscore the role of creativity in modulating both the expressive and physical dimensions of human movement. Moreover, this research demonstrates the potential of inertial sensor technology not only to capture kinematic patterns but also to provide insight into the deeper layers of human artistic and cognitive processes. Full article

Low back pain (LBP) remains one of the most prevalent and disabling musculoskeletal conditions globally, with profound social, economic, and healthcare implications. The rising incidence and chronic nature of LBP highlight the need for more objective, personalized, and effective approaches to assessment and rehabilitation. In this context, bioengineering has emerged as a transformative field, offering novel tools and methodologies that enhance the understanding and management of LBP. This narrative review examines current bioengineering applications in both diagnostic and therapeutic domains. For assessment, technologies such as wearable inertial sensors, three-dimensional motion capture systems, surface electromyography, and biomechanical modeling provide real-time, quantitative insights into posture, movement patterns, and muscle activity. On the therapeutic front, innovations including robotic exoskeletons, neuromuscular electrical stimulation, virtual reality-based rehabilitation, and tele-rehabilitation platforms are increasingly being integrated into multimodal treatment protocols. These technologies support precision medicine by tailoring interventions to each patient’s biomechanical and functional profile. Furthermore, the incorporation of artificial intelligence into clinical workflows enables automated data analysis, predictive modeling, and decision support systems, while future directions such as digital twin technology hold promise for personalized simulation and outcome forecasting. While these advancements are promising, further validation in large-scale, real-world settings is required to ensure safety, efficacy, and equitable accessibility. Ultimately, bioengineering provides a multidimensional, data-driven framework that has the potential to significantly improve the assessment, rehabilitation, and overall management of LBP. Full article

With the rapid development of smart manufacturing, the optimization of real-time monitoring in operating procedures has become a crucial issue in modern industry. Traditional disassembly and assembly (D/A) work, relying on human experience and visual inspection, lacks immediacy and a quantitative basis, further affecting operating quality and efficiency. This study aims to develop a thin-film force sensor and an inertial measurement unit (IMU)-integrated wearable device for monitoring and analyzing operators’ behavioral characteristics during D/A tasks. First, by having operators wear self-made smart gloves and 17 IMU sensors, the work tables with three different heights are equipped with a mechatronics machine for the D/A experiment. Common D/A motions are designed into the experiment. Several subjects are invited to execute the standardized operating procedure, with upper limbs used to collect data on operators’ hand gestures and movements. Then, the measured data are applied to verify the performance measure functional best path of machine D/A. The results reveal that the system could effectively identify various D/A motions as well as observe operators’ force difference and motion mode, which, through the theory of performance indicator optimization and the verification of data analysis, could provide a reference for the best path planning, D/A sequence, and work table height design in the machine D/A process. The optimal workbench height for a standing operator is 5 to 10 cm above their elbow height. Performing assembly and disassembly tasks at this optimal height can help the operator save between 14.3933% and 35.2579% of physical effort. Such outcomes could aid in D/A behavior monitoring in industry, worker training, and operational optimization, as well as expand the application to instant feedback design for automation and smartization in a smart factory. Full article

The assessment of upper limb (UL) movement patterns plays a critical role in the rehabilitation of individuals with motor impairments resulting from neuromotor disorders, which significantly affect essential activities of daily living (ADLs) such as drinking and eating. However, conventional clinical evaluation methods often lack objective and quantitative insights into the biomechanics of movement. To enable accurate identification of pathological patterns, it is first necessary to establish normative biomechanical and electrophysiological benchmarks in healthy individuals. In this study, a previously developed, low-cost, wearable, and portable prototype device was employed to objectively assess UL movement. The system, specifically designed for clinical applicability, integrates surface electromyography (EMG) sensors and an inertial measurement unit (IMU) to capture muscle activity and kinematic data, respectively. Thirty healthy participants were recruited to perform standardized drinking and eating tasks. The analysis focused on characterizing muscle activation patterns and joint range of motion during different task phases. Results revealed consistent variations in muscle contraction and joint kinematics, allowing the identification of distinct activation profiles for key shoulder muscles. The findings contribute to the establishment of a normative dataset that can serve as a reference for the assessment of clinical populations. Such data are essential for informing rehabilitation strategies and developing predictive models of UL function during ADLs in individuals with neuromotor disorders. Full article

This paper addresses the problem of recognising and classifying hand movements to control an upper-limb exoskeleton. To solve this problem, a multisensory system based on the fusion of data from electromyography (EMG) sensors, inertial measurement units (IMUs), and virtual reality (VR) trackers is proposed, which provides highly accurate detection of users’ movements. Signal preprocessing (noise filtering, segmentation, normalisation) and feature extraction were performed to generate input data for regression and classification models. Various machine learning algorithms are used to recognise motor activity, ranging from classical algorithms (logistic regression, k-nearest neighbors, decision trees) and ensemble methods (random forest, AdaBoost, eXtreme Gradient Boosting, stacking, voting) to deep neural networks, including convolutional neural networks (CNNs), gated recurrent units (GRUs), and transformers. The algorithm for integrating machine learning models into the exoskeleton control system is considered. In experiments aimed at abandoning proprietary tracking systems (VR trackers), absolute position regression was performed using data from IMU sensors with 14 regression algorithms: The random forest ensemble provided the best accuracy (mean absolute error = 0.0022 metres). The task of classifying activity categories out of nine types is considered below. Ablation analysis showed that IMU and VR trackers produce a sufficient informative minimum, while adding EMG also introduces noise, which degrades the performance of simpler models but is successfully compensated for by deep networks. In the classification task using all signals, the maximum result (99.2%) was obtained on Transformer; the fully connected neural network generated slightly worse results (98.4%). When using only IMU data, fully connected neural network, Transformer, and CNN–GRU networks provide 100% accuracy. Experimental results confirm the effectiveness of the proposed architectures for motor activity classification, as well as the use of a multi-sensor approach that allows one to compensate for the limitations of individual types of sensors. The obtained results make it possible to continue research in this direction towards the creation of control systems for upper exoskeletons, including those used in rehabilitation and virtual simulation systems. Full article

Objectives: The aim of this study was to evaluate the impact of four working heights on lumbar biomechanics during wall construction tasks, focusing on work-related musculoskeletal disorders (WMSDs). Methods: Fifteen young male participants performed simulated mortar-spreading and bricklaying tasks while actual body movements were recorded using Inertial Measurement Unit (IMU) sensors. Muscle activities of the lumbar erector spinae (ES), quadratus lumborum (QL), multifidus (MF), gluteus maximus (GM), and iliopsoas (IL) were estimated using a 3D musculoskeletal (MSK) model and measured via surface electromyography (sEMG). The analysis of variance (ANOVA) test was conducted to identify the significant differences in muscle activities across four working heights (i.e., foot, knee, waist, and shoulder). Results: Findings showed that working at foot-level height resulted in the highest muscle activity (7.6% to 40.6% increase), particularly in the ES and QL muscles, indicating an increased risk of WMSDs. The activities of the ES, MF, and GM muscles were statistically significant across both tasks and all working heights (p < 0.01). Conclusions: Both MSK and sEMG analyses indicated significantly lower muscle activities at knee and waist heights, suggesting these as the best working positions (47 cm to 107 cm) for minimizing the risk of WMSDs. Conversely, working at foot and shoulder heights was identified as a significant risk factor for WMSDs. Additionally, the similar trends observed between MSK simulations and sEMG data suggest that MSK modeling can effectively substitute for sEMG in future studies. These findings provide valuable insights into ergonomic work positioning to reduce WMSD risks among wall construction workers. Full article

Accurate estimation of upper-limb joint angles is essential in biomechanics, rehabilitation, and wearable robotics. While inertial measurement units (IMUs) offer portability and flexibility, systems requiring multiple inertial sensors can be intrusive and complex to deploy. In contrast, optical motion capture (MoCap) systems provide precise tracking but are constrained to controlled laboratory environments. This study presents a deep learning-based approach for estimating shoulder and elbow joint angles using only three IMU sensors positioned on the chest and both wrists, validated against reference angles obtained from a MoCap system. The input data includes Euler angles, accelerometer, and gyroscope data, synchronized and segmented into sliding windows. Two recurrent neural network architectures, Convolutional Neural Network with Long-short Term Memory (CNN-LSTM) and Bidirectional LSTM (BLSTM), were trained and evaluated using identical conditions. The CNN component enabled the LSTM to extract spatial features that enhance sequential pattern learning, improving angle reconstruction. Both models achieved accurate estimation performance: CNN-LSTM yielded lower Mean Absolute Error (MAE) in smooth trajectories, while BLSTM provided smoother predictions but underestimated some peak movements, especially in the primary axes of rotation. These findings support the development of scalable, deep learning-based wearable systems and contribute to future applications in clinical assessment, sports performance analysis, and human motion research. Full article