IMU Data-Driven and PCA-Based Approach to Establish Quantifiable and Practically Applicable Measures for V2 Technique Elements in Cross-Country Skiing.

Quantifying movement coordination in cross-country (XC) skiing, specifically the technique with its elemental forms, is challenging. Particularly, this applies when trying to establish a bidirectional transfer between scientific theory and practical experts' knowledge as expressed, for example, in ski instruction curricula. The objective of this study was to translate 14 curricula-informed distinct elements of the V2 ski-skating technique (horizontal and vertical posture, lateral tilt, head position, upper body rotation, arm swing, shoulder abduction, elbow flexion, hand and leg distance, plantar flexion, ski set-down, leg push-off, and gliding phase) into plausible, valid and applicable measures to make the technique training process more quantifiable and scientifically grounded. Inertial measurement unit (IMU) data of 10 highly experienced XC skiers who demonstrated the technique elements by two extreme forms each (e.g., anterior versus posterior positioning for the horizontal posture) were recorded. Element-specific principal component analyses (PCAs)-driven by the variance produced by the technique extremes-resulted in movement components that express quantifiable measures of the underlying technique elements. Ten measures were found to be sensitive in distinguishing between the inputted extreme variations using statistical parametric mapping (SPM), whereas for four elements the SPM did not detect differences (lateral tilt, plantar flexion, ski set-down, and leg push-off). Applicability of the established technique measures was determined based on quantifying individual techniques through them. The study introduces a novel approach to quantitatively assess V2 ski-skating technique, which might help to enhance technique feedback and bridge the communication gap that often exists between practitioners and scientists.

Kinematic characteristics of elbow joint range of motion in elite Chinese freestyle swimmers with elbow pain during dry-land simulations of swimming strokes.

The aim of this study was to obtain quantitative data on elbow joint ROM in elite freestyle swimmers with EP in China. Of the 50 elite freestyle swimmers recruited, 41 completed all measurements during dry-land swimming stroke simulations. Elbow joint angle, velocity, and acceleration were measured using inertial measurement units. The RMSE/D was calculated to determine the elbow joint ROM deviation. Joint angle (3.33 ∘-42.96 ∘), angular velocity (-364.15 to 245.69 ∘/s), and angular acceleration (-7051.80 to 1465.35 ∘/s2) were significantly different between the critical pain and healthy. The probability distributions of joint angle (15.47 ∘ ±14.54 ∘), angular velocity (2.41 ∘ ±111.06 ∘/s), and angular acceleration (1.93 ± 2222.6 ∘/s2) in the slight pain group were significantly different betweenhealthy and critical pain. The RMSE/D distributions of angular velocity (28.3%) and acceleration (21.48%) in the critical pain deviated from the healthy. The peak value-RMSE/D matrix model obtained proved that elbow ROM significantly differed between the elite freestyle swimmers with EP and the healthy. Angular velocity and acceleration indicate the weakness and negative influence of kinematics on patients with EP. Thus, Potential solutions are to constantly optimise freestyle swimming techniques and strengthen the arm muscles.

Parameterization of Biomechanical Variables through Inertial Measurement Units (IMUs) in Occasional Healthy Runners.

Running is one of the most popular sports practiced today and biomechanical variables are fundamental to understanding it. The main objectives of this study are to describe kinetic, kinematic, and spatiotemporal variables measured using four inertial measurement units (IMUs) in runners during treadmill running, investigate the relationships between these variables, and describe differences associated with different data sampling and averaging strategies. A total of 22 healthy recreational runners (M age = 28 ± 5.57 yrs) participated in treadmill measurements, running at their preferred speed (M = 10.1 ± 1.9 km/h) with a set-up of four IMUs placed on tibias and the lumbar area. Raw data was processed and analysed over selections spanning 30 s, 30 steps and 1 step. Very strong positive associations were obtained between the same family variables in all selections. The temporal variables were inversely associated with the step rate variable in the selection of 30 s and 30 steps of data. There were moderate associations between kinetic (forces) and kinematic (displacement) variables. There were no significant differences between the biomechanics variables in any selection. Our results suggest that a 4-IMU set-up, as presented in this study, is a viable approach for parameterization of the biomechanical variables in running, and also that there are no significant differences in the biomechanical variables studied independently, if we select data from 30 s, 30 steps or 1 step for processing and analysis. These results can assist in the methodological aspects of protocol design in future running research.

Exploring the Influence of BMI on Gait Metrics: A Comprehensive Analysis of Spatiotemporal Parameters and Stability Indicators.

BACKGROUND: Gait analysis is a vital tool for evaluating overall health and predicting outcomes such as mortality and cognitive decline. This study explores how normal and obese BMI categories impact gait dynamics, addressing gaps in understanding the effect of body composition on specific gait parameters. RESEARCH QUESTION: The primary objective is to investigate differences in spatiotemporal gait parameters-specifically, gait speed, step length, cadence, and double support time-between normal and obese BMI groups to understand the effects of obesity on gait. METHODS: This observational case-control study analyzed spatiotemporal gait metrics from 163 participants, using inertial measurement units (IMUs) to collect data on various gait parameters. Statistical analyses explored the relationship between BMI categories and these metrics. RESULTS: No significant differences were found in gait speed, cadence, stride duration, or double support time between the normal and obese groups. However, significant differences were identified in age, hypertension prevalence, balance problems, and the incidence of falls, emphasizing the complex effects of obesity on factors influencing gait stability. SIGNIFICANCE: This study contributes to our understanding of obesity's impact on gait by highlighting the need to consider associated health and stability parameters. These findings prompt a re-evaluation of how BMI is integrated into clinical gait assessments and emphasize the necessity for personalized healthcare strategies. This research highlights the importance of future studies with larger, more diverse populations and a wider array of biomechanical measures to dissect the relationship between BMI, body composition, and gait dynamics.

Quantifying Hand Motion Complexity in Simulated Sailing Using Inertial Sensors.

The control of hand movement during sailing is important for performance. To quantify the amount of regularity and the unpredictability of hand fluctuations during the task, the mathematical algorithm Approximate Entropy (ApEn) of the hand acceleration can be used. Approximate Entropy is a mathematical algorithm that depends on the combination of two input parameters including (1) the length of the sequences to be compared (m), and (2) the tolerance threshold for accepting similar patterns between two segments (r). The aim of this study is to identify the proper combinations of 'm' and 'r' parameter values for ApEn measurement in the hand movement acceleration data during sailing. Inertial Measurement Units (IMUs) recorded acceleration data for both the mainsail (non-dominant) and tiller (dominant) hands across the X-, Y-, and Z-axes, as well as vector magnitude. ApEn values were computed for 24 parameter combinations, with 'm' ranging from 2 to 5 and 'r' from 0.10 to 0.50. The analysis revealed significant differences in acceleration ApEn regularity between the two hands, particularly along the Z-axis, where the mainsail hand exhibited higher entropy values (p = 0.000673), indicating greater acceleration complexity and unpredictability. In contrast, the tiller hand displayed more stable and predictable acceleration patterns, with lower ApEn values. ANOVA results confirmed that parameter 'm' had a significant effect on acceleration complexity for both hands, highlighting differing motor control demands between the mainsail and tiller hands. These findings demonstrate the utility of IMU sensors and ApEn in detecting nuanced variations in acceleration dynamics during sailing tasks. This research contributes to the understanding of hand-specific acceleration patterns in sailing and provides a foundation for further studies on adaptive sailing techniques and motor control strategies for both novice and expert sailors.

Integrating sEMG and IMU Sensors in an e-Textile Smart Vest for Forward Posture Monitoring: First Steps.

Currently, the market for wearable devices is expanding, with a growing trend towards the use of these devices for continuous-monitoring applications. Among these, real-time posture monitoring and assessment stands out as a crucial application given the rising prevalence of conditions like forward head posture (FHP). This paper proposes a wearable device that combines the acquisition of electromyographic signals from the cervical region with inertial data from inertial measurement units (IMUs) to assess the occurrence of FHP. To improve electronics integration and wearability, e-textiles are explored for the development of surface electrodes and conductive tracks that connect the different electronic modules. Tensile strength and abrasion tests of 22 samples consisting of textile electrodes and conductive tracks produced with three fiber types (two from Shieldex and one from Imbut) were conducted. Imbut's Elitex fiber outperformed Shieldex's fibers in both tests. The developed surface electromyography (sEMG) acquisition hardware and textile electrodes were also tested and benchmarked against an electromyography (EMG) gold standard in dynamic and isometric conditions, with results showing slightly better root mean square error (RMSE) values (for 4 × 2 textile electrodes (10.02%) in comparison to commercial Ag/AgCl electrodes (11.11%). The posture monitoring module was also validated in terms of joint angle estimation and presented an overall error of 4.77° for a controlled angular velocity of 40°/s as benchmarked against a UR10 robotic arm.

Using Video Technology and AI within Parkinson's Disease Free-Living Fall Risk Assessment.

Falls are a major concern for people with Parkinson's disease (PwPD), but accurately assessing real-world fall risk beyond the clinic is challenging. Contemporary technologies could enable the capture of objective and high-resolution data to better inform fall risk through measurement of everyday factors (e.g., obstacles) that contribute to falls. Wearable inertial measurement units (IMUs) capture objective high-resolution walking/gait data in all environments but are limited by not providing absolute clarity on contextual information (i.e., obstacles) that could greatly influence how gait is interpreted. Video-based data could compliment IMU-based data for a comprehensive free-living fall risk assessment. The objective of this study was twofold. First, pilot work was conducted to propose a novel artificial intelligence (AI) algorithm for use with wearable video-based eye-tracking glasses to compliment IMU gait data in order to better inform free-living fall risk in PwPD. The suggested approach (based on a fine-tuned You Only Look Once version 8 (YOLOv8) object detection algorithm) can accurately detect and contextualize objects (mAP50 = 0.81) in the environment while also providing insights into where the PwPD is looking, which could better inform fall risk. Second, we investigated the perceptions of PwPD via a focus group discussion regarding the adoption of video technologies and AI during their everyday lives to better inform their own fall risk. This second aspect of the study is important as, traditionally, there may be clinical and patient apprehension due to ethical and privacy concerns on the use of wearable cameras to capture real-world video. Thematic content analysis was used to analyse transcripts and develop core themes and categories. Here, PwPD agreed on ergonomically designed wearable video-based glasses as an optimal mode of video data capture, ensuring discreteness and negating any public stigma on the use of research-style equipment. PwPD also emphasized the need for control in AI-assisted data processing to uphold privacy, which could overcome concerns with the adoption of video to better inform IMU-based gait and free-living fall risk. Contemporary technologies (wearable video glasses and AI) can provide a holistic approach to fall risk that PwPD recognise as helpful and safe to use.

Opal Actigraphy (Activity and Sleep) Measures Compared to ActiGraph: A Validation Study.

Physical activity and sleep monitoring in daily life provide vital information to track health status and physical fitness. The aim of this study was to establish concurrent validity for the new Opal Actigraphy solution in relation to the widely used ActiGraph GT9X for measuring physical activity from accelerometry epic counts (sedentary to vigorous levels) and sleep periods in daily life. Twenty participants (age 56 + 22 years) wore two wearable devices on each wrist for 7 days and nights, recording 3-D accelerations at 30 Hz. Bland-Altman plots and intraclass correlation coefficients (ICCs) assessed validity (agreement) and test-retest reliability between ActiGraph and Opal Actigraphy sleep durations and activity levels, as well as between the two different versions of the ActiGraph. ICCs showed excellent reliability for physical activity measures and moderate-to-excellent reliability for sleep measures between Opal versus Actigraph GT9X and between GT3X versus GT9X. Bland-Altman plots and mean absolute percentage error (MAPE) also show a comparable performance (within 10%) between Opal and ActiGraph and between the two ActiGraph monitors across activity and sleep measures. In conclusion, physical activity and sleep measures using Opal Actigraphy demonstrate performance comparable to that of ActiGraph, supporting concurrent validation. Opal Actigraphy can be used to quantify activity and monitor sleep patterns in research and clinical studies.

BioMAT: An Open-Source Biomechanics Multi-Activity Transformer for Joint Kinematic Predictions Using Wearable Sensors.

Through wearable sensors and deep learning techniques, biomechanical analysis can reach beyond the lab for clinical and sporting applications. Transformers, a class of recent deep learning models, have become widely used in state-of-the-art artificial intelligence research due to their superior performance in various natural language processing and computer vision tasks. The performance of transformer models has not yet been investigated in biomechanics applications. In this study, we introduce a Biomechanical Multi-activity Transformer-based model, BioMAT, for the estimation of joint kinematics from streaming signals of multiple inertia measurement units (IMUs) using a publicly available dataset. This dataset includes IMU signals and the corresponding sagittal plane kinematics of the hip, knee, and ankle joints during multiple activities of daily living. We evaluated the model's performance and generalizability and compared it against a convolutional neural network long short-term model, a bidirectional long short-term model, and multi-linear regression across different ambulation tasks including level ground walking (LW), ramp ascent (RA), ramp descent (RD), stair ascent (SA), and stair descent (SD). To investigate the effect of different activity datasets on prediction accuracy, we compared the performance of a universal model trained on all activities against task-specific models trained on individual tasks. When the models were tested on three unseen subjects' data, BioMAT outperformed the benchmark models with an average root mean square error (RMSE) of 5.5 ± 0.5°, and normalized RMSE of 6.8 ± 0.3° across all three joints and all activities. A unified BioMAT model demonstrated superior performance compared to individual task-specific models across four of five activities. The RMSE values from the universal model for LW, RA, RD, SA, and SD activities were 5.0 ± 1.5°, 6.2 ± 1.1°, 5.8 ± 1.1°, 5.3 ± 1.6°, and 5.2 ± 0.7° while these values for task-specific models were, 5.3 ± 2.1°, 6.7 ± 2.0°, 6.9 ± 2.2°, 4.9 ± 1.4°, and 5.6 ± 1.3°, respectively. Overall, BioMAT accurately estimated joint kinematics relative to previous machine learning algorithms across different activities directly from the sequence of IMUs signals instead of time-normalized gait cycle data.

Acceleration-Based Estimation of Vertical Ground Reaction Forces during Running: A Comparison of Methods across Running Speeds, Surfaces, and Foot Strike Patterns.

Twenty-seven methods of estimating vertical ground reaction force first peak, loading rate, second peak, average, and/or time series from a single wearable accelerometer worn on the shank or approximate center of mass during running were compared. Force estimation errors were quantified for 74 participants across different running surfaces, speeds, and foot strike angles and biases, repeatability coefficients, and limits of agreement were modeled with linear mixed effects to quantify the accuracy, reliability, and precision. Several methods accurately and reliably estimated the first peak and loading rate, however, none could do so precisely (the limits of agreement exceeded ±65% of target values). Thus, we do not recommend first peak or loading rate estimation from accelerometers with the methods currently available. In contrast, the second peak, average, and time series could all be estimated accurately, reliably, and precisely with several different methods. Of these, we recommend the 'Pogson' methods due to their accuracy, reliability, and precision as well as their stability across surfaces, speeds, and foot strike angles.

Three-Dimensional Human Pose Estimation from Sparse IMUs through Temporal Encoder and Regression Decoder.

Three-dimensional (3D) pose estimation has been widely used in many three-dimensional human motion analysis applications, where inertia-based path estimation is gradually being adopted. Systems based on commercial inertial measurement units (IMUs) usually rely on dense and complex wearable sensors and time-consuming calibration, causing intrusions to the subject and hindering free body movement. The sparse IMUs-based method has drawn research attention recently. Existing sparse IMUs-based three-dimensional pose estimation methods use neural networks to obtain human poses from temporal feature information. However, these methods still suffer from issues, such as body shaking, body tilt, and movement ambiguity. This paper presents an approach to improve three-dimensional human pose estimation by fusing temporal and spatial features. Based on a multistage encoder-decoder network, a temporal convolutional encoder and human kinematics regression decoder were designed. The final three-dimensional pose was predicted from the temporal feature information and human kinematic feature information. Extensive experiments were conducted on two benchmark datasets for three-dimensional human pose estimation. Compared to state-of-the-art methods, the mean per joint position error was decreased by 13.6% and 19.4% on the total capture and DIP-IMU datasets, respectively. The quantitative comparison demonstrates that the proposed temporal information and human kinematic topology can improve pose accuracy.

A Multi-Modal Under-Sensorized Wearable System for Optimal Kinematic and Muscular Tracking of Human Upper Limb Motion.

Wearable sensing solutions have emerged as a promising paradigm for monitoring human musculoskeletal state in an unobtrusive way. To increase the deployability of these systems, considerations related to cost reduction and enhanced form factor and wearability tend to discourage the number of sensors in use. In our previous work, we provided a theoretical solution to the problem of jointly reconstructing the entire muscular-kinematic state of the upper limb, when only a limited amount of optimally retrieved sensory data are available. However, the effective implementation of these methods in a physical, under-sensorized wearable has never been attempted before. In this work, we propose to bridge this gap by presenting an under-sensorized system based on inertial measurement units (IMUs) and surface electromyography (sEMG) electrodes for the reconstruction of the upper limb musculoskeletal state, focusing on the minimization of the sensors' number. We found that, relying on two IMUs only and eight sEMG sensors, we can conjointly reconstruct all 17 degrees of freedom (five joints, twelve muscles) of the upper limb musculoskeletal state, yielding a median normalized RMS error of 8.5% on the non-measured joints and 2.5% on the non-measured muscles.

The effect of training schedule and playing positions on training loads and game demands in professional handball players.

In this research, we aimed to (1) describe the differences in internal and external load between playing positions and (2) characterize the training demands of the days before competitive events for professional handball players. Fifteen players (5 wings, 2 centre backs, 4 backs, and 2 pivots) were equipped with a local positioning system device during training and 11 official matches. External (total distance, high-speed running, player load) and internal loads (rating of perceived exertion) were computed. Substantial differences were recorded between the external load variables depending on each playing position and depending on whether it was a training day (high-speed running: effect size (ES) ≥ 2.07; player load: ES ≥ 1.89) or a match (total distance: ES ≥ 1.27; high-speed running: ES ≥ 1.42; player load: ES ≥ 1.33). Differences in internal load were not substantial. The rating of perceived exertion, at this competitive level, does not seem to discriminate the differences registered in the external load, probably due to the degree of adaptation to the specific effort of these players. The large differences observed in external load variables should be used to tailor practices and better adjust the training demands in professional handball settings.

Assessing Cerebellar Disorders with Wearable Inertial Sensor Data Using Time-Frequency and Autoregressive Hidden Markov Model Approaches.

Wearable sensor data is relatively easily collected and provides direct measurements of movement that can be used to develop useful behavioral biomarkers. Sensitive and specific behavioral biomarkers for neurodegenerative diseases are critical to supporting early detection, drug development efforts, and targeted treatments. In this paper, we use autoregressive hidden Markov models and a time-frequency approach to create meaningful quantitative descriptions of behavioral characteristics of cerebellar ataxias from wearable inertial sensor data gathered during movement. We create a flexible and descriptive set of features derived from accelerometer and gyroscope data collected from wearable sensors worn while participants perform clinical assessment tasks, and use these data to estimate disease status and severity. A short period of data collection (<5 min) yields enough information to effectively separate patients with ataxia from healthy controls with very high accuracy, to separate ataxia from other neurodegenerative diseases such as Parkinson's disease, and to provide estimates of disease severity.

Quantitative Characterization of Motor Control during Gait in Dravet Syndrome Using Wearable Sensors: A Preliminary Study.

Dravet syndrome (DS) is a rare and severe form of genetic epilepsy characterized by cognitive and behavioural impairments and progressive gait deterioration. The characterization of gait parameters in DS needs efficient, non-invasive quantification. The aim of the present study is to apply nonlinear indexes calculated from inertial measurements to describe the dynamics of DS gait. Twenty participants (7 M, age 9-33 years) diagnosed with DS were enrolled. Three wearable inertial measurement units (OPAL, Apdm, Portland, OR, USA; Miniwave, Cometa s.r.l., Italy) were attached to the lower back and ankles and 3D acceleration and angular velocity were acquired while participants walked back and forth along a straight path. Segmental kinematics were acquired by means of stereophotogrammetry (SMART, BTS). Community functioning data were collected using the functional independence measure (FIM). Mean velocity and step width were calculated from stereophotogrammetric data; fundamental frequency, harmonic ratio, recurrence quantification analysis, and multiscale entropy (τ = 1...6) indexes along anteroposterior (AP), mediolateral (ML), and vertical (V) axes were calculated from trunk acceleration. Results were compared to a reference age-matched control group (112 subjects, 6-25 years old). All nonlinear indexes show a disruption of the cyclic pattern of the centre of mass in the sagittal plane, quantitatively supporting the clinical observation of ataxic gait. Indexes in the ML direction were less altered, suggesting the efficacy of the compensatory strategy (widening the base of support). Nonlinear indexes correlated significantly with functional scores (i.e., FIM and speed), confirming their effectiveness in capturing clinically meaningful biomarkers of gait.

Assessing Respiratory Activity by Using IMUs: Modeling and Validation.

This study aimed to explore novel inertial measurement unit (IMU)-based strategies to estimate respiratory parameters in healthy adults lying on a bed while breathing normally. During the experimental sessions, the kinematics of the chest wall were contemporaneously collected through both a network of 9 IMUs and a set of 45 uniformly distributed reflective markers. All inertial kinematics were analyzed to identify a minimum set of signals and IMUs whose linear combination best matched the tidal volume measured by optoelectronic plethysmography. The resulting models were finally tuned and validated through a leave-one-out cross-validation approach to assess the extent to which they could accurately estimate a set of respiratory parameters related to three trunk compartments. The adopted methodological approach allowed us to identify two different models. The first, referred to as Model 1, relies on the 3D acceleration measured by three IMUs located on the abdominal compartment and on the lower costal margin. The second, referred to as Model 2, relies on only one component of the acceleration measured by two IMUs located on the abdominal compartment. Both models can accurately estimate the respiratory rate (relative error < 1.5%). Conversely, the duration of the respiratory phases and the tidal volume can be more accurately assessed by Model 2 (relative error < 5%) and Model 1 (relative error < 5%), respectively. We further discuss possible approaches to overcome limitations and improve the overall accuracy of the proposed approach.

Inertial Measurement Unit Sensors in Assistive Technologies for Visually Impaired People, a Review.

A diverse array of assistive technologies have been developed to help Visually Impaired People (VIP) face many basic daily autonomy challenges. Inertial measurement unit sensors, on the other hand, have been used for navigation, guidance, and localization but especially for full body motion tracking due to their low cost and miniaturization, which have allowed the estimation of kinematic parameters and biomechanical analysis for different field of applications. The aim of this work was to present a comprehensive approach of assistive technologies for VIP that include inertial sensors as input, producing results on the comprehension of technical characteristics of the inertial sensors, the methodologies applied, and their specific role in each developed system. The results show that there are just a few inertial sensor-based systems. However, these sensors provide essential information when combined with optical sensors and radio signals for navigation and special application fields. The discussion includes new avenues of research, missing elements, and usability analysis, since a limitation evidenced in the selected articles is the lack of user-centered designs. Finally, regarding application fields, it has been highlighted that a gap exists in the literature regarding aids for rehabilitation and biomechanical analysis of VIP. Most of the findings are focused on navigation and obstacle detection, and this should be considered for future applications.

The Accuracy and Precision of Gait Spatio-Temporal Parameters Extracted from an Instrumented Sock during Treadmill and Overground Walking in Healthy Subjects and Patients with a Foot Impairment Secondary to Psoriatic Arthritis.

The objectives of this study were to assess the accuracy and precision of a system combining an IMU-instrumented sock and a validated algorithm for the estimation of the spatio-temporal parameters of gait. A total of 25 healthy participants (HP) and 21 patients with foot impairments secondary to psoriatic arthritis (PsA) performed treadmill walking at three different speeds and overground walking at a comfortable speed. HP performed the assessment over two sessions. The proposed system's estimations of cadence (CAD), gait cycle duration (GCD), gait speed (GS), and stride length (SL) obtained for treadmill walking were validated versus those estimated with a motion capture system. The system was also compared with a well-established multi-IMU-based system for treadmill and overground walking. The results showed a good agreement between the motion capture system and the IMU-instrumented sock in estimating the spatio-temporal parameters during the treadmill walking at normal and fast speeds for both HP and PsA participants. The accuracy of GS and SL obtained from the IMU-instrumented sock was better compared to the established multi-IMU-based system in both groups. The precision (inter-session reliability) of the gait parameter estimations obtained from the IMU-instrumented sock was good to excellent for overground walking and treadmill walking at fast speeds, but moderate-to-good for slow and normal treadmill walking. The proposed IMU-instrumented sock offers a novel form factor addressing the wearability issues of IMUs and could potentially be used to measure spatio-temporal parameters under clinical conditions and free-living conditions.

Estimation of Human Center of Mass Position through the Inertial Sensors-Based Methods in Postural Tasks: An Accuracy Evaluation.

The estimation of the body's center of mass (CoM) trajectory is typically obtained using force platforms, or optoelectronic systems (OS), bounding the assessment inside a laboratory setting. The use of magneto-inertial measurement units (MIMUs) allows for more ecological evaluations, and previous studies proposed methods based on either a single sensor or a sensors' network. In this study, we compared the accuracy of two methods based on MIMUs. Body CoM was estimated during six postural tasks performed by 15 healthy subjects, using data collected by a single sensor on the pelvis (Strapdown Integration Method, SDI), and seven sensors on the pelvis and lower limbs (Biomechanical Model, BM). The accuracy of the two methods was compared in terms of RMSE and estimation of posturographic parameters, using an OS as reference. The RMSE of the SDI was lower in tasks with little or no oscillations, while the BM outperformed in tasks with greater CoM displacement. Moreover, higher correlation coefficients were obtained between the posturographic parameters obtained with the BM and the OS. Our findings showed that the estimation of CoM displacement based on MIMU was reasonably accurate, and the use of the inertial sensors network methods should be preferred to estimate the kinematic parameters.

Fusion Models for Generalized Classification of Multi-Axial Human Movement: Validation in Sport Performance.

We introduce a set of input models for fusing information from ensembles of wearable sensors supporting human performance and telemedicine. Veracity is demonstrated in action classification related to sport, specifically strikes in boxing and taekwondo. Four input models, formulated to be compatible with a broad range of classifiers, are introduced and two diverse classifiers, dynamic time warping (DTW) and convolutional neural networks (CNNs) are implemented in conjunction with the input models. Seven classification models fusing information at the input-level, output-level, and a combination of both are formulated. Action classification for 18 boxing punches and 24 taekwondo kicks demonstrate our fusion classifiers outperform the best DTW and CNN uni-axial classifiers. Furthermore, although DTW is ostensibly an ideal choice for human movements experiencing non-linear variations, our results demonstrate deep learning fusion classifiers outperform DTW. This is a novel finding given that CNNs are normally designed for multi-dimensional data and do not specifically compensate for non-linear variations within signal classes. The generalized formulation enables subject-specific movement classification in a feature-blind fashion with trivial computational expense for trained CNNs. A commercial boxing system, 'Corner', has been produced for real-world mass-market use based on this investigation providing a basis for future telemedicine translation.