Classifying running speed conditions using a single wearable sensor: Optimal segmentation and feature extraction methods

Use of subject-specific models to detect fatigue-related changes in running biomechanics: a random forest approach

Running biomechanics are affected by fatiguing or prolonged runs. However, no evidence to date has conclusively linked this effect to running-related injury (RRI) development or performance implications. Previous investigations using subject-specific models in running have demonstrated higher accuracy than group-based models, however, this has been infrequently applied to fatigue. In this study, two experiments were conducted to determine whether subject-specific models outperformed group-based models to classify running biomechanics during non-fatigued and fatigued conditions. In the first experiment, 16 participants performed four treadmill runs at or around the maximal lactate steady state. In the second experiment, nine participants performed five prolonged runs using commercial wearable devices. For each experiment, two segments were extracted from each trial from early and late in the run. For each participant, a random forest model was applied with a leave-one-run-out cross-validation to classify between the early (non-fatigued) and late (fatigued) segments. Additionally, group-based classifiers with a leave-one-subject-out cross validation were constructed. For experiment 1, mean classification accuracies for the single-subject and group-based classifiers were 68.2 ± 8.2% and 57.0 ± 8.9%, respectively. For experiment 2, mean classification accuracies for the single-subject and group-based classifiers were 68.9 ± 17.1% and 61.5 ± 11.7%, respectively. Variable importance rankings were consistent within participants, but these rankings differed from each participant to those of the group. Although the classification accuracies were relatively low, these findings highlight the advantage of subject-specific classifiers to detect changes in running biomechanics with fatigue and indicate the potential of using big data and wearable technology approaches in future research to determine possible connections between biomechanics and RRI.

Using Lower Limb Wearable Sensors to Identify Gait Modalities: A Machine-Learning-Based Approach

Real-world gait analysis can aid in clinical assessments and influence related interventions, free from the restrictions of a laboratory setting. Using individual accelerometers, we aimed to use a simple machine learning method to quantify the performance of the discrimination between three self-selected cyclical locomotion types using accelerometers placed at frequently referenced attachment locations. Thirty-five participants walked along a 10 m walkway at three different speeds. Triaxial accelerometers were attached to the sacrum, thighs and shanks. Slabs of magnitude, three-second-long accelerometer data were transformed into two-dimensional Fourier spectra. Principal component analysis was undertaken for data reduction and feature selection, followed by discriminant function analysis for classification. Accuracy was quantified by calculating scalar accounting for the distances between the three centroids and the scatter of each category's cloud. The algorithm could successfully discriminate between gait modalities with 91% accuracy at the sacrum, 90% at the shanks and 87% at the thighs. Modalities were discriminated with high accuracy in all three sensor locations, where the most accurate location was the sacrum. Future research will focus on optimising the data processing of information from sensor locations that are advantageous for practical reasons, e.g., shank for prosthetic and orthotic devices.

Wearables for Running Gait Analysis: A Systematic Review

BACKGROUND

Running gait assessment has traditionally been performed using subjective observation or expensive laboratory-based objective technologies, such as three-dimensional motion capture or force plates. However, recent developments in wearable devices allow for continuous monitoring and analysis of running mechanics in any environment. Objective measurement of running gait is an important (clinical) tool for injury assessment and provides measures that can be used to enhance performance.

OBJECTIVES

We aimed to systematically review the available literature investigating how wearable technology is being used for running gait analysis in adults.

METHODS

A systematic search of the literature was conducted in the following scientific databases: PubMed, Scopus, Web of Science and SPORTDiscus. Information was extracted from each included article regarding the type of study, participants, protocol, wearable device(s), main outcomes/measures, analysis and key findings.

RESULTS

A total of 131 articles were reviewed: 56 investigated the validity of wearable technology, 22 examined the reliability and 77 focused on applied use. Most studies used inertial measurement units (n = 62) [i.e. a combination of accelerometers, gyroscopes and magnetometers in a single unit] or solely accelerometers (n = 40), with one using gyroscopes alone and 31 using pressure sensors. On average, studies used one wearable device to examine running gait. Wearable locations were distributed among the shank, shoe and waist. The mean number of participants was 26 (± 27), with an average age of 28.3 (± 7.0) years. Most studies took place indoors (n = 93), using a treadmill (n = 62), with the main aims seeking to identify running gait outcomes or investigate the effects of injury, fatigue, intrinsic factors (e.g. age, sex, morphology) or footwear on running gait outcomes. Generally, wearables were found to be valid and reliable tools for assessing running gait compared to reference standards.

CONCLUSIONS

This comprehensive review highlighted that most studies that have examined running gait using wearable sensors have done so with young adult recreational runners, using one inertial measurement unit sensor, with participants running on a treadmill and reporting outcomes of ground contact time, stride length, stride frequency and tibial acceleration. Future studies are required to obtain consensus regarding terminology, protocols for testing validity and the reliability of devices and suitability of gait outcomes.

CLINICAL TRIAL REGISTRATION

CRD42021235527.

Investigating walking speed variability of young adults in the real world

BACKGROUND

Walking speed strongly correlates with health outcomes, making accurate assessment essential for clinical evaluations. However, assessments tend to be conducted over short distances, often in a laboratory or clinical setting, and may not capture natural walking behavior. To address this gap, the following questions are investigated in this work: Is walking speed significantly influenced by the continuity and duration of a walking bout? Can preferred walking speed be inferred by grouping walking bouts using duration and continuity?

METHODS

We collected two weeks of continuous data from fifteen healthy young adults using a thigh-worn accelerometer and a heart rate monitor. Walking strides were identified and grouped into walking periods. We quantified the duration and the continuity of each walking period. Continuity is used to parameterize changes in stepping rate related to pauses during a bout of walking. Finally, we analyzed the influence of duration and continuity on estimates of stride speed, and examined how the distribution of walking speed varies depending on different walking modes (defined by duration and continuity).

RESULTS

We found that continuity and duration can be used to explain some of the variability in real-world walking speed (p<0.001). Speeds estimated from long continuous walks with many strides (42% of all recorded strides) had the lowest standard deviation. Walking speed during these bouts was 1.41ms^-1 (SD = 0.26ms^-1).

SIGNIFICANCE

Walking behavior in the real world is largely variable. Features of real-world walks, like duration and continuity, can be used to explain some of the variability observed in walking speed. As such, we recommend using long continuous walks to confidently isolate the preferred walking behavior of an individual.

Recent Machine Learning Progress in Lower Limb Running Biomechanics With Wearable Technology: A Systematic Review

With the emergence of wearable technology and machine learning approaches, gait monitoring in real-time is attracting interest from the sports biomechanics community. This study presents a systematic review of machine learning approaches in running biomechanics using wearable sensors. Electronic databases were retrieved in PubMed, Web of Science, SPORTDiscus, Scopus, IEEE Xplore, and ScienceDirect. A total of 4,068 articles were identified via electronic databases. Twenty-four articles that met the eligibility criteria after article screening were included in this systematic review. The range of quality scores of the included studies is from 0.78 to 1.00, with 40% of articles recruiting participant numbers between 20 and 50. The number of inertial measurement unit (IMU) placed on the lower limbs varied from 1 to 5, mainly in the pelvis, thigh, distal tibia, and foot. Deep learning algorithms occupied 57% of total machine learning approaches. Convolutional neural networks (CNN) were the most frequently used deep learning algorithm. However, the validation process for machine learning models was lacking in some studies and should be given more attention in future research. The deep learning model combining multiple CNN and recurrent neural networks (RNN) was observed to extract different running features from the wearable sensors and presents a growing trend in running biomechanics.

Between-Day Reliability of Commonly Used IMU Features during a Fatiguing Run and the Effect of Speed

The purpose of this study was to determine if fatigue-related changes in biomechanics derived from an inertial measurement unit (IMU) placed at the center of mass (CoM) are reliable day-to-day. Sixteen runners performed two runs at maximal lactate steady state (MLSS) on a treadmill, one run 5% above MLSS speed, and one run 5% below MLSS speed while wearing a CoM-mounted IMU. Trials were performed to volitional exhaustion or a specified termination time. IMU features were derived from each axis and the resultant. Feature means were calculated for each subject during non-fatigued and fatigued states. Comparisons were performed between the two trials at MLSS and between all four trials. The only significant fatigue state × trial interaction was the 25th percentile of the results when comparing all trials. There were no main effects for trial for either comparison method. There were main effects for fatigue state for most features in both comparison methods. Reliability, measured by an intraclass coefficient (ICC), was good-to-excellent for most features. These results suggest that fatigue-related changes in biomechanics derived from a CoM-mounted IMU are reliable day-to-day when participants ran at or around MLSS and are not significantly affected by slight deviations in speed.

Is This the Real Life, or Is This Just Laboratory? A Scoping Review of IMU-Based Running Gait Analysis

Inertial measurement units (IMUs) can be used to monitor running biomechanics in real-world settings, but IMUs are often used within a laboratory. The purpose of this scoping review was to describe how IMUs are used to record running biomechanics in both laboratory and real-world conditions. We included peer-reviewed journal articles that used IMUs to assess gait quality during running. We extracted data on running conditions (indoor/outdoor, surface, speed, and distance), device type and location, metrics, participants, and purpose and study design. A total of 231 studies were included. Most (72%) studies were conducted indoors; and in 67% of all studies, the analyzed distance was only one step or stride or <200 m. The most common device type and location combination was a triaxial accelerometer on the shank (18% of device and location combinations). The most common analyzed metric was vertical/axial magnitude, which was reported in 64% of all studies. Most studies (56%) included recreational runners. For the past 20 years, studies using IMUs to record running biomechanics have mainly been conducted indoors, on a treadmill, at prescribed speeds, and over small distances. We suggest that future studies should move out of the lab to less controlled and more real-world environments.

Wearable Multi-Functional Sensing Technology for Healthcare Smart Detection

In recent years, considerable research efforts have been devoted to the development of wearable multi-functional sensing technology to fulfill the requirements of healthcare smart detection, and much progress has been achieved. Due to the appealing characteristics of flexibility, stretchability and long-term stability, the sensors have been used in a wide range of applications, such as respiration monitoring, pulse wave detection, gait pattern analysis, etc. Wearable sensors based on single mechanisms are usually capable of sensing only one physiological or motion signal. In order to measure, record and analyze comprehensive physical conditions, it is indispensable to explore the wearable sensors based on hybrid mechanisms and realize the integration of multiple smart functions. Herein, we have summarized various working mechanisms (resistive, capacitive, triboelectric, piezoelectric, thermo-electric, pyroelectric) and hybrid mechanisms that are incorporated into wearable sensors. More importantly, to make wearable sensors work persistently, it is meaningful to combine flexible power units and wearable sensors and form a self-powered system. This article also emphasizes the utility of self-powered wearable sensors from the perspective of mechanisms, and gives applications. Furthermore, we discuss the emerging materials and structures that are applied to achieve high sensitivity. In the end, we present perspectives on the outlooks of wearable multi-functional sensing technology.

A support vector machine algorithm can successfully classify running ability when trained with wearable sensor data from anatomical locations typical of consumer technology

Greater understanding of differences in technique between runners may allow more beneficial feedback related to improving performance and decreasing injury risk. The purpose of this study was to develop and test a support vector machine classifier, which could automatically differentiate running technique between experienced and novice participants using only wearable sensor data. Three-dimensional linear accelerations and angular velocities were collected from six wearable sensors secured to current common smart device locations. Cross-validation was used to test the classification accuracy of models trained with a variety of combinations of sensor locations, with participants running at different speeds. Average classification accuracies ranged from 71.3% to 98.4% across the sensor combinations and running speeds tested. Models trained with only a single sensor location still showed effective classification. With the models trained with only upper arm data achieving an average accuracy of 96.4% across all tested running speeds. A post-hoc comparison of biomechanical variables between the two subgroups showed significant differences in upper body biomechanics throughout the stride. Both the methodology used to perform the classifications and the biomechanical differences identified could prove useful when aiming to shift a novice runner's technique towards movement patterns more akin to those with greater experience.

Machine learning approaches applied in spinal pain research

The purpose of this narrative review is to provide a critical reflection of how analytical machine learning approaches could provide the platform to harness variability of patient presentation to enhance clinical prediction. The review includes a summary of current knowledge on the physiological adaptations present in people with spinal pain. We discuss how contemporary evidence highlights the importance of not relying on single features when characterizing patients given the variability of physiological adaptations present in people with spinal pain. The advantages and disadvantages of current analytical strategies in contemporary basic science and epidemiological research are reviewed and we consider how analytical machine learning approaches could provide the platform to harness the variability of patient presentations to enhance clinical prediction of pain persistence or recurrence. We propose that machine learning techniques can be leveraged to translate a potentially heterogeneous set of variables into clinically useful information with the potential to enhance patient management.

Feature selection to classify lameness using a smartphone-based inertial measurement unit

Background and objectives

Gait can be severely affected by pain, muscle weakness, and aging resulting in lameness. Despite the high incidence of lameness, there are no studies on the features that are useful for classifying lameness patterns. Therefore, we aimed to identify features of high importance for classifying population differences in lameness patterns using an inertial measurement unit mounted above the sacral region.

Methods

Features computed exhaustively for multidimensional time series consisting of three-axis angular velocities and three-axis acceleration were carefully selected using the Benjamini–Yekutieli procedure, and multiclass classification was performed using LightGBM (Microsoft Corp., Redmond, WA, USA). We calculated the relative importance of the features that contributed to the classification task in machine learning.

Results

The most important feature was found to be the absolute value of the Fourier coefficients of the second frequency calculated by the one-dimensional discrete Fourier transform for real input. This was determined by the fast Fourier transformation algorithm using data of a single gait cycle of the yaw angular velocity of the pelvic region.

Conclusions

Using an inertial measurement unit worn over the sacral region, we determined a set of features of high importance for classifying differences in lameness patterns based on different factors. This completely new set of indicators can be used to advance the understanding of lameness.

Workload a-WEAR-ness: Monitoring Workload in Team Sports With Wearable Technology. A Scoping Review

OBJECTIVES

To (1) identify the wearable devices and associated metrics used to monitor workload and assess injury risk, (2) describe the situations in which workload was monitored using wearable technology (including sports, purpose of the analysis, location and duration of monitoring, and athlete characteristics), and (3) evaluate the quality of evidence that workload monitoring can inform injury prevention.

DESIGN

Scoping review.

LITERATURE SEARCH

We searched the CINAHL, Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews, Embase, HealthSTAR, MEDLINE, PsycINFO, SPORTDiscus, and Web of Science databases.

STUDY SELECTION CRITERIA

We included all studies that used wearable devices (eg, heart rate monitor, inertial measurement units, global positioning system) to monitor athlete workload in a team sport setting.

DATA SYNTHESIS

We provided visualizations that represented the workload metrics reported, sensors used, sports investigated, athlete characteristics, and the duration of monitoring.

RESULTS

The 407 included studies focused on team ball sports (67% soccer, rugby, or Australian football), male athletes (81% of studies), elite or professional level of competition (74% of studies), and young adults (69% of studies included athletes aged between 20 and 28 years). Thirty-six studies of 7 sports investigated the association between workload measured with wearable devices and injury.

CONCLUSION

Distance-based metrics derived from global positioning system units were common for monitoring workload and are frequently used to assess injury risk. Workload monitoring studies have focused on specific populations (eg, elite male soccer players in Europe and elite male rugby and Australian football players in Oceania). Different injury definitions and reported workload metrics and poor study quality impeded conclusions regarding the relationship between workload and injury. J Orthop Sports Phys Ther 2020;50(10):549-563. doi:10.2519/jospt.2020.9753.

Evaluating upper limb function after stroke using the free-living accelerometer data

Accelerometer devices are becoming efficient tools in clinical studies for automatically measuring the activities of daily living. Such data provides a time series describing activity level at every second and displays a subject’s activity pattern throughout a day. However, the analysis of such data is very challenging due to the large number of observations produced each second and the variability among subjects. The purpose of this study is to develop efficient statistical analysis techniques for predicting the recovery level of the upper limb function after stroke based on the free-living accelerometer data. We propose to use a Gaussian Mixture Model (GMM)-based method for clustering and extracting new features to capture the information contained in the raw data. A nonlinear mixed effects model with Gaussian Process prior for the random effects is developed as the predictive model for evaluating the recovery level of the upper limb function. Results of applying to the accelerometer data for patients after stroke are presented.

Gait Variability Using Waist- and Ankle-Worn Inertial Measurement Units in Healthy Older Adults

Gait variability observed in step duration is predictive of impending adverse health outcomes among apparently healthy older adults and could potentially be evaluated using wearable sensors (inertial measurement units, IMU). The purpose of the present study was to establish the reliability and concurrent validity of gait variability and complexity evaluated with a waist and an ankle-worn IMU. Seventeen women (age 74.8 (SD 44) years) and 10 men (73.7 (4.1) years) attended two laboratory measurement sessions a week apart. Their stride duration variability was concurrently evaluated based on a continuous 3 min walk using a force plate and a waist- and an ankle-worn IMU. Their gait complexity (multiscale sample entropy) was evaluated from the waist-worn IMU. The force plate indicated excellent stride duration variability reliability (intra-class correlation coefficient, ICC = 0.90), whereas fair to good reliability (ICC = 0.47 to 0.66) was observed from the IMUs. The IMUs exhibited poor to excellent concurrent validity in stride duration variability compared to the force plate (ICC = 0.22 to 0.93). A good to excellent reliability was observed for gait complexity in most coarseness scales (ICC = 0.60 to 0.82). A reasonable congruence with the force plate-measured stride duration variability was observed on many coarseness scales (correlation coefficient = 0.38 to 0.83). In conclusion, waist-worn IMU entropy estimates may provide a feasible indicator of gait variability among community-dwelling ambulatory older adults.

Interpretable machine learning models for classifying low back pain status using functional physiological variables

PURPOSE

To evaluate the predictive performance of statistical models which distinguishes different low back pain (LBP) sub-types and healthy controls, using as input predictors the time-varying signals of electromyographic and kinematic variables, collected during low-load lifting.

METHODS

Motion capture with electromyography (EMG) assessment was performed on 49 participants [healthy control (con) = 16, remission LBP (rmLBP) = 16, current LBP (LBP) = 17], whilst performing a low-load lifting task, to extract a total of 40 predictors (kinematic and electromyographic variables). Three statistical models were developed using functional data boosting (FDboost), for binary classification of LBP statuses (model 1: con vs. LBP; model 2: con vs. rmLBP; model 3: rmLBP vs. LBP). After removing collinear predictors (i.e. a correlation of > 0.7 with other predictors) and inclusion of the covariate sex, 31 predictors were included for fitting model 1, 31 predictors for model 2, and 32 predictors for model 3.

RESULTS

Seven EMG predictors were selected in model 1 (area under the receiver operator curve [AUC] of 90.4%), nine predictors in model 2 (AUC of 91.2%), and seven predictors in model 3 (AUC of 96.7%). The most influential predictor was the biceps femoris muscle (peak [Formula: see text]  = 0.047) in model 1, the deltoid muscle (peak [Formula: see text] =  0.052) in model 2, and the iliocostalis muscle (peak [Formula: see text] =  0.16) in model 3.

CONCLUSION

The ability to transform time-varying physiological differences into clinical differences could be used in future prospective prognostic research to identify the dominant movement impairments that drive the increased risk. These slides can be retrieved under Electronic Supplementary Material.

New Considerations for Collecting Biomechanical Data Using Wearable Sensors: The Effect of Different Running Environments

Traditionally, running biomechanics analyses have been conducted using 3D motion capture during treadmill or indoor overground running. However, most runners complete their runs outdoors. Since changes in running terrain have been shown to influence running gait mechanics, the purpose of this study was to use a machine learning approach to objectively determine relevant accelerometer-based features to discriminate between running patterns in different environments and determine the generalizability of observed differences in running patterns. Center of mass accelerations were recorded for recreational runners in treadmill-only (n = 28) and sidewalk-only (n = 25) environments, and an independent group (n = 16) ran in both treadmill and sidewalk environments. A feature selection algorithm was used to develop a training dataset from treadmill-only and sidewalk-only running. A binary support vector machine model was trained to classify treadmill and sidewalk running. Classification accuracy was determined using 10-fold cross-validation of the training dataset and an independent testing dataset from the runners that ran in both environments. Nine features related to the consistency and variability of center of mass accelerations were selected. Specifically, there was greater ratio of vertical acceleration during treadmill running and a greater ratio of anterior-posterior acceleration during sidewalk running in both the training and testing dataset. Step and stride regularity were significantly greater in the treadmill condition for the vertical axis in both the training and testing dataset, and in the medial-lateral axis for the testing dataset. During sidewalk running, there was significantly greater variability in the magnitude of the vertical and anterior-posterior accelerations for both datasets. The classification accuracy based on 10-fold cross-validation of the training dataset (M = 93.17%, SD = 2.43%) was greater than the classification accuracy of the independent testing dataset (M = 83.81%, SD = 3.39%). This approach could be utilized in future analyses to identify relevant differences in running patterns using wearable technology.

Classifying neck pain status using scalar and functional biomechanical variables - development of a method using functional data boosting

BACKGROUND

Individuals with neck pain have different movement and muscular activation (collectively termed as biomechanical variables) patterns compared to healthy individuals. Incorporating biomechanical variables as covariates into prognostic models is challenging due to the high dimensionality of the data.

RESEARCH QUESTION

What is the classification performance of neck pain status of a statistical model which uses both scalar and functional biomechanical covariates?

METHODS

Motion capture with electromyography assessment on the sternocleidomastoid, splenius cervicis, erector spinae, was performed on 21 healthy and 26 individuals with neck pain during walking over three gait conditions (rectilinear, curvilinear clockwise (CW) and counterclockwise (CCW)). After removing highly collinear variables, 94 covariates across the three conditions were used to classify neck pain status using functional data boosting (FDboost).

RESULTS

Two functional covariates trunk lateral flexion angle during CCW gait, and trunk flexion angle during CW gait; and a scalar covariate, hip jerk index during CCW gait were selected. The model achieved an estimated AUC of 80.8 %. For hip jerk index, an increase in hip jerk index by one unit increased the log odds of being in the neck pain group by 0.37. A 1° increase in trunk lateral flexion angle throughout gait alone reduced the probability of being in the neck pain group from 0.5 to 0.15. A 1° increase in trunk flexion angle throughout gait alone increased the probability of being in the neck pain group from 0.5 to 0.9.

SIGNIFICANCE

Interpreting the physiological significance of the extracted covariates, with other biomechanical variables, suggests that individuals with neck pain performed curvilinear walking using a stiffer strategy, compared to controls; and this increased the risk of being in the neck pain group. FDboost can produce clinically interpretable models with complex high dimensional data and could be used in future prognostic modelling studies in neck pain research.

Step by Step Towards Effective Human Activity Recognition: A Balance between Energy Consumption and Latency in Health and Wellbeing Applications

Human activity recognition (HAR) is a classification process that is used for recognizing human motions. A comprehensive review of currently considered approaches in each stage of HAR, as well as the influence of each HAR stage on energy consumption and latency is presented in this paper. It highlights various methods for the optimization of energy consumption and latency in each stage of HAR that has been used in literature and was analyzed in order to provide direction for the implementation of HAR in health and wellbeing applications. This paper analyses if and how each stage of the HAR process affects energy consumption and latency. It shows that data collection and filtering and data segmentation and classification stand out as key stages in achieving a balance between energy consumption and latency. Since latency is only critical for real-time HAR applications, the energy consumption of sensors and devices stands out as a key challenge for HAR implementation in health and wellbeing applications. Most of the approaches in overcoming challenges related to HAR implementation take place in the data collection, filtering and classification stages, while the data segmentation stage needs further exploration. Finally, this paper recommends a balance between energy consumption and latency for HAR in health and wellbeing applications, which takes into account the context and health of the target population.

New Considerations for Wearable Technology Data: Changes in Running Biomechanics During a Marathon

The purpose of this study was to use wearable technology data to quantify alterations in subject-specific running patterns throughout a marathon race and to determine if runners could be clustered into subgroups based on similar trends in running gait alterations throughout the marathon. Using a wearable sensor, data were collected for cadence, braking, bounce, pelvic rotation, pelvic drop, and ground contact time for 27 runners. A composite index was calculated based on the "typical" data (4-14 km) for each runner and evaluated for 14 individual 2-km sections thereafter to detect "atypical" data (ie, higher indices). A cluster analysis assigned all runners to a subgroup based on similar trends in running alterations. Results indicated that the indices became significantly higher starting at 20 to 22 km. Cluster 1 exhibited lower indices than cluster 2 throughout the marathon, and the only significant difference in characteristics between clusters was that cluster 1 had a lower age-grade performance score than cluster 2. In summary, this study presented a novel method to investigate the effects of fatigue on running biomechanics using wearable technology in a real-world setting. Recreational runners with higher age-grade performance scores had less atypical running patterns throughout the marathon compared with runners with lower age-grade performance scores.

Single Inertial Sensor-Based Neural Networks to Estimate COM-COP Inclination Angle During Walking

A biomechanical understanding of gait stability is needed to reduce falling risk. As a typical parameter, the COM-COP (center of mass-center of pressure) inclination angle (IA) could provide valuable insight into postural control and balance recovery ability. In this study, an artificial neural network (ANN) model was developed to estimate COM-COP IA based on signals using an inertial sensor. Also, we evaluated how different types of ANN and the cutoff frequency of the low-pass filter applied to input signals could affect the accuracy of the model. An inertial measurement unit (IMU) including an accelerometer, gyroscope, and magnetometer sensors was fabricated as a prototype. The COM-COP IA was calculated using a 3D motion analysis system including force plates. In order to predict the COM-COP IA, a feed-forward ANN and long-short term memory (LSTM) network was developed. As a result, the feed-forward ANN showed a relative root-mean-square error (rRMSE) of 15% while the LSTM showed an improved accuracy of 9% rRMSE. Additionally, the LSTM displayed a stable accuracy regardless of the cutoff frequency of the filter applied to the input signals. This study showed that estimating the COM-COP IA was possible with a cheap inertial sensor system. Furthermore, the neural network models in this study can be implemented in systems to monitor the balancing ability of the elderly or patients with impaired balancing ability.

Effect of running speed on temporal and frequency indicators from wearable MEMS accelerometers

Amplified by the development of new technologies, the interest in personal performance has been growing over the last years. Acceleration has proved to be an easy variable to collect, and was addressed in several works. However, few of them evaluate the effect of running speed on relevant indicators. The influence of the sensors location on the measurement is rarely studied as well. This study is dedicated to investigating the effect of running speed on acceleration measured at three different positions on 18 volunteers. All participants were equipped with three inertial measurement units: on the dorsal surface of the right foot (Fo), at the centre of gravity of the tibia (Ti), at the L4-L5 lumbar (Lu). The test was performed on a treadmill at nine randomised speeds between 8 and 18 km/h. Ten accelerometric variables were calculated. Linear regressions were used to calculate speed from the indicators calculated on (Lu), (Ti), (Fo). Indicators associated to signal energy were highly correlated with speed (r2>0.90). Median frequency appears to be affected by the frequency resolution. Finally, the measurement points closest to the impact zone result in the most correlated indicators.

Automated Accelerometer-Based Gait Event Detection During Multiple Running Conditions

The identification of the initial contact (IC) and toe off (TO) events are crucial components of running gait analyses. To evaluate running gait in real-world settings, robust gait event detection algorithms that are based on signals from wearable sensors are needed. In this study, algorithms for identifying gait events were developed for accelerometers that were placed on the foot and low back and validated against a gold standard force plate gait event detection method. These algorithms were automated to enable the processing of large quantities of data by accommodating variability in running patterns. An evaluation of the accuracy of the algorithms was done by comparing the magnitude and variability of the difference between the back and foot methods in different running conditions, including different speeds, foot strike patterns, and outdoor running surfaces. The results show the magnitude and variability of the back-foot difference was consistent across running conditions, suggesting that the gait event detection algorithms can be used in a variety of settings. As wearable technology allows for running gait analyses to move outside of the laboratory, the use of automated accelerometer-based gait event detection methods may be helpful in the real-time evaluation of running patterns in real world conditions.

Continuous Analysis of Running Mechanics by Means of an Integrated INS/GPS Device

This paper describes a single body-mounted sensor that integrates accelerometers, gyroscopes, compasses, barometers, a GPS receiver, and a methodology to process the data for biomechanical studies. The sensor and its data processing system can accurately compute the speed, acceleration, angular velocity, and angular orientation at an output rate of 400 Hz and has the ability to collect large volumes of ecologically-valid data. The system also segments steps and computes metrics for each step. We analyzed the sensitivity of these metrics to changing the start time of the gait cycle. Along with traditional metrics, such as cadence, speed, step length, and vertical oscillation, this system estimates ground contact time and ground reaction forces using machine learning techniques. This equipment is less expensive and cumbersome than the currently used alternatives: Optical tracking systems, in-shoe pressure measurement systems, and force plates. Another advantage, compared to existing methods, is that natural movement is not impeded at the expense of measurement accuracy. The proposed technology could be applied to different sports and activities, including walking, running, motion disorder diagnosis, and geriatric studies. In this paper, we present the results of tests in which the system performed real-time estimation of some parameters of walking and running which are relevant to biomechanical research. Contact time and ground reaction forces computed by the neural network were found to be as accurate as those obtained by an in-shoe pressure measurement system.

Subject-specific and group-based running pattern classification using a single wearable sensor

The objective of this study was to determine whether subject-specific or group-based models provided better classification accuracy to identify changes in biomechanical running gait patterns across different inclination conditions. The classification process was based on measurements from a single wearable sensor using a total of 41,780 strides from eleven recreational runners while running in real-world and uncontrolled environment. Biomechanical variables included pelvic drop, ground contact time, braking, vertical oscillation of pelvis, pelvic rotation, and cadence were recorded during running on three inclination grades: downhill, -2° to -7°; level, -0.2° to +0.2°; and uphill, +2° to +7°. An ensemble and non-linear machine learning algorithm, random forest (RF), was used to classify inclination condition and determine the importance of each of the biomechanical variables. Classification accuracy was determined for subject-specific and group-based RF models. The mean classification accuracy of all subject-specific RF models was 86.29%, while group-based classification accuracy was 76.17%. Braking was identified as the most important variable for all the runners using the group-based model and for most of the runners based on a subject-specific models. In addition, individual runners used different strategies across different inclination conditions and the ranked order of variable importance was unique for each runner. These results demonstrate that subject-specific models can better characterize changes in gait biomechanical patterns compared to a more traditional group-based approach.

Validity of a novel method to measure vertical oscillation during running using a depth camera

Recent advancements in low-cost depth cameras may provide a clinically accessible alternative to conventional three-dimensional (3D) multi-camera motion capture systems for gait analysis. However, there remains a lack of information on the validity of clinically relevant running gait parameters such as vertical oscillation (VO). The purpose of this study was to assess the validity of measures of VO during running gait using raw depth data, in comparison to a 3D multi-camera motion capture system. Sixteen healthy adults ran on a treadmill at a standard speed of 2.7 m/s. The VO of their running gait was simultaneously collected from raw depth data (Microsoft Kinect v2) and 3D marker data (Vicon multi-camera motion capture system). The agreement between the VO measures obtained from the two systems was assessed using a Bland-Altman plot with 95% limits of agreement (LOA), a Pearson's correlation coefficient (r), and a Lin's concordance correlation coefficient (r_c). The depth data from the Kinect v2 demonstrated excellent results across all measures of validity (r = 0.97; r_c = 0.97; 95% LOA = -8.0 mm - 8.7 mm), with an average absolute error and percent error of 3.7 (2.1) mm and 4.0 (2.0)%, respectively. The findings of this study have demonstrated the ability of a low cost depth camera and a novel tracking method to accurately measure VO in running gait.

Wearable Sensor Data to Track Subject-Specific Movement Patterns Related to Clinical Outcomes Using a Machine Learning Approach

Wearable sensors can provide detailed information on human movement but the clinical impact of this information remains limited. We propose a machine learning approach, using wearable sensor data, to identify subject-specific changes in gait patterns related to improvements in clinical outcomes. Eight patients with knee osteoarthritis (OA) completed two gait trials before and one following an exercise intervention. Wearable sensor data (e.g., 3-dimensional (3D) linear accelerations) were collected from a sensor located near the lower back, lateral thigh and lateral shank during level treadmill walking at a preferred speed. Wearable sensor data from the 2 pre-intervention gait trials were used to define each individual’s typical movement pattern using a one-class support vector machine (OCSVM). The percentage of strides defined as outliers, based on the pre-intervention gait data and the OCSVM, were used to define the overall change in an individual’s movement pattern. The correlation between the change in movement patterns following the intervention (i.e., percentage of outliers) and improvement in self-reported clinical outcomes (e.g., pain and function) was assessed using a Spearman rank correlation. The number of outliers observed post-intervention exhibited a large association (ρ = 0.78) with improvements in self-reported clinical outcomes. These findings demonstrate a proof-of-concept and a novel methodological approach for integrating machine learning and wearable sensor data. This approach provides an objective and evidence-informed way to understand clinically important changes in human movement patterns in response to exercise therapy.