Gait detection in children with and without hemiplegia using single-axis wearable gyroscopes

Automatic gait events detection with inertial measurement units: healthy subjects and moderate to severe impaired patients

BACKGROUND

Recently, the use of inertial measurement units (IMUs) in quantitative gait analysis has been widely developed in clinical practice. Numerous methods have been developed for the automatic detection of gait events (GEs). While many of them have achieved high levels of efficiency in healthy subjects, detecting GEs in highly degraded gait from moderate to severely impaired patients remains a challenge. In this paper, we aim to present a method for improving GE detection from IMU recordings in such cases.

METHODS

We recorded 10-meter gait IMU signals from 13 healthy subjects, 29 patients with multiple sclerosis, and 21 patients with post-stroke equino varus foot. An instrumented mat was used as the gold standard. Our method detects GEs from filtered acceleration free from gravity and gyration signals. Firstly, we use autocorrelation and pattern detection techniques to identify a reference stride pattern. Next, we apply multiparametric Dynamic Time Warping to annotate this pattern from a model stride, in order to detect all GEs in the signal.

RESULTS

We analyzed 16,819 GEs recorded from healthy subjects and achieved an F1-score of 100%, with a median absolute error of 8 ms (IQR [3-13] ms). In multiple sclerosis and equino varus foot cohorts, we analyzed 6067 and 8951 GEs, respectively, with F1-scores of 99.4% and 96.3%, and median absolute errors of 18 ms (IQR [8-39] ms) and 26 ms (IQR [12-50] ms).

CONCLUSIONS

Our results are consistent with the state of the art for healthy subjects and demonstrate a good accuracy in GEs detection for pathological patients. Therefore, our proposed method provides an efficient way to detect GEs from IMU signals, even in degraded gaits. However, it should be evaluated in each cohort before being used to ensure its reliability.

Robust and adaptive terrain classification and gait event detection system

Real-time gait event detection (GED) system can be utilized for gait analysis and tracking fitness activities. GED for various types of terrains (e.g., stair-walk, uneven surfaces, etc.) is still an open research problem. This study presents an inertial sensor-based approach for real-time GED system that works for diverse terrains in an uncontrolled environment. The GED system classifies three types of terrains, i.e., flat-walk, stair-ascend and stair-descend, with an average classification accuracy of 99%. It also accurately detects various gait events, including, toe-strike, heel-rise, toe-off, and heel-strike. It is computationally efficient, implemented on a low-cost microcontroller, works in real-time and can be used in portable rehabilitation devices for use in dynamic environments.

The Use of Embedded IMU Insoles to Assess Gait Parameters: A Validation and Test-Retest Reliability Study

Wireless wearable insoles are interesting tools to collect gait parameters during daily life activities. However, studies have to be performed specifically for each type of insoles on a big data set to validate the measurement in ecological situations. This study aims to assess the criterion validity and test-retest reliability of gait parameters from wearable insoles compared to motion capture system. Gait of 30 healthy participants was recorded using DSPro® insoles and a motion capture system during overground and treadmill walking at three different speeds. Criterion validity and test-retest reliability of spatio-temporal parameters were estimated with an intraclass correlation coefficient (ICC). For both systems, reliability was found higher than 0.70 for all variables (p < 0.001) except for minimum toe clearance (ICC < 0.50) with motion capture system during overground walking. Regardless of speed and condition of walking, Speed, Cadence, Stride Length, Stride Time and Stance Time variables were validated (ICC > 0.90; p < 0.001). During walking on treadmill, loading time was not validated during slow speed (ICC < 0.70). This study highlights good criterion validity and test-retest reliability of spatiotemporal gait parameters measurement using wearable insoles and opens a new possibility to improve care management of patients using clinical gait analysis in daily life activities.

A Real-Time Gait Phase Detection Method Based on BiLSTM-Attention Model

Real-time gait phase detection is essential to achieve accurate and stable walking assistance in intelligent rehabilitation training for patients with motor disorders. This study proposed an efficient real-time detection method to detect three gait phases (loading response, stance, and swing) based on a bidirectional long short-term memory network with an attention layer (BiLSTM-Attention). We validated our method on a public dataset where eight healthy subjects' data during treadmill walking were employed. A single inertial measurement unit (IMU) was attached to the shank to measure the sagittal plane acceleration of the lower leg and the angular velocity around the central lateral axis. These data were transposed and segmented into data sequences based on labels using a sliding window method. The data from 8 participants were divided into the training, validation, and test sets (5:1:2). Results showed the average recognition accuracy of the proposed model on new subjects was 97.40% with an average time delay of 15.7±10.1ms, showing the method's potential to be applied for practice use.

Classification of gait phases based on a machine learning approach using muscle synergy

The accurate detection of the gait phase is crucial for monitoring and diagnosing neurological and musculoskeletal disorders and for the precise control of lower limb assistive devices. In studying locomotion mode identification and rehabilitation of neurological disorders, the concept of modular organization, which involves the co-activation of muscle groups to generate various motor behaviors, has proven to be useful. This study aimed to investigate whether muscle synergy features could provide a more accurate and robust classification of gait events compared to traditional features such as time-domain and wavelet features. For this purpose, eight healthy individuals participated in this study, and wireless electromyography sensors were attached to four muscles in each lower extremity to measure electromyography (EMG) signals during walking. EMG signals were segmented and labeled as 2-class (stance and swing) and 3-class (weight acceptance, single limb support, and limb advancement) gait phases. Non-negative matrix factorization (NNMF) was used to identify specific muscle groups that contribute to gait and to provide an analysis of the functional organization of the movement system. Gait phases were classified using four different machine learning algorithms: decision tree (DT), k-nearest neighbors (KNN), support vector machine (SVM), and neural network (NN). The results showed that the muscle synergy features had a better classification accuracy than the other EMG features. This finding supported the hypothesis that muscle synergy enables accurate gait phase classification. Overall, the study presents a novel approach to gait analysis and highlights the potential of muscle synergy as a tool for gait phase detection.

A scoping review on recent trends in wearable sensors to analyze gait in people with stroke: From sensor placement to validation against gold-standard equipment

The purpose of the review is to evaluate wearable sensor placement, their impact and validation of wearable sensors on analyzing gait, primarily the postural instability in people with stroke. Databases, namely PubMed, Cochrane, SpringerLink, and IEEE Xplore were searched to identify related articles published since January 2005. The authors have selected the articles by considering patient characteristics, intervention details, and outcome measurements by following the priorly set inclusion and exclusion criteria. From a total of 1077 articles, 142 were included in this study and classified into functional fields, namely postural stability (PS) assessments, physical activity monitoring (PA), gait pattern classification (GPC), and foot drop correction (FDC). The review covers the types of wearable sensors, their placement, and their performance in terms of reliability and validity. When employing a single wearable sensor, the pelvis and foot were the most used locations for detecting gait asymmetry and kinetic parameters, respectively. Multiple Inertial Measurement Units placed at different body parts were effectively used to estimate postural stability and gait pattern. This review article has compared results of placement of sensors at different locations helping researchers and clinicians to identify the best possible placement for sensors to measure specific kinematic and kinetic parameters in persons with stroke.

Predicting vertical ground reaction forces from 3D accelerometry using reservoir computers leads to accurate gait event detection

Accelerometers are low-cost measurement devices that can readily be used outside the lab. However, determining isolated gait events from accelerometer signals, especially foot-off events during running, is an open problem. We outline a two-step approach where machine learning serves to predict vertical ground reaction forces from accelerometer signals, followed by force-based event detection. We collected shank accelerometer signals and ground reaction forces from 21 adults during comfortable walking and running on an instrumented treadmill. We trained one common reservoir computer using segmented data using both walking and running data. Despite being trained on just a small number of strides, this reservoir computer predicted vertical ground reaction forces in continuous gait with high quality. The subsequent foot contact and foot off event detection proved highly accurate when compared to the gold standard based on co-registered ground reaction forces. Our proof-of-concept illustrates the capacity of combining accelerometry with machine learning for detecting isolated gait events irrespective of mode of locomotion.

Concurrent validity, test-retest reliability, and sensitivity of a PostureRite system measurement on dynamic postural sway and risk of fall in cerebral palsy

BACKGROUND: Accurately diagnosing dynamic postural sway (DPS) is essential for effective and sustainable intervention in children with cerebral palsy (CP). We developed an accurate, inexpensive, and wearable DPS measurement system to measure DPS accurately and consistently during walking and functional activities of daily living. OBJECTIVE: We investigated the validity and reliability of this PostureRite system in children with CP, and the link between PostureRite and clinical measures including gross motor function measure (GMFM), pediatric balance scale (PBS), and fall efficacy scale (FES). METHODS: Twenty-one participants were categorized as follows: 11 healthy adults (3 females, mean age, 25.00±1.00 years) and 10 children with CP (mean age, 11.10±6.28 years). We determined the concurrent validity of PostureRite by comparing DPS data to the gold standard accelerometer measurement results. We determined test-retest reliability by measuring DPS data on three occasions at 2-h intervals. We assessed PostureRite measurement sensitivity to ascertain differences between healthy children and children with CP DPS measurements. RESULTS: Random and mixed intraclass correlation coefficients (ICC2,k and ICC3,k) were obtained; an independent T-test was performed (P < 0.05). Concurrent validity analysis showed a good relationship between the gold standard accelerometer and PostureRite (ICC2,k = 0.973, P < 0.05). Test-retest reliability demonstrated a good relationship across the three repeated measures of the DPS data (ICC3,k = 0.816–0.924, P < 0.05). Independent T-test revealed a significant difference in DPS data between healthy adults and children with CP (P < 0.05). CONCLUSIONS: We developed a portable, wireless, and affordable PostureRite system to measure DPS during gross motor function associated with daily activity and participation, and established the concurrent validity, test-retest reliability as sensitivity, and clinical relevance by comparing the DPS obtained from the participants with and without CP.

Intelligent wearable system with accurate detection of abnormal gait and timely cueing for mobility enhancement of people with Parkinson's disease

Previously reported wearable systems for people with Parkinson's disease (PD) have been focused on the detection of abnormal gait. They suffered from limited accuracy, large latency, poor durability, comfort, and convenience for daily use. Herewith we report an intelligent wearable system (IWS) that can accurately detect abnormal gait in real-time and provide timely cueing for PD patients. The system features novel sensitive, comfortable and durable plantar pressure sensing insoles with a highly compressed data set, an accurate and fast gait algorithm, and wirelessly controlled timely sensory cueing devices. A total of 29 PD patients participated in the first phase without cueing for developing processes of the algorithm, which achieved an accuracy of over 97% for off-line detection of freezing of gait (FoG). In the second phase with cueing, the evaluation of the whole system was conducted with 16 PD subjects via trial and a questionnaire survey. This system demonstrated an accuracy of 94% for real-time detection of FoG and a mean latency of 0.37 s between the onset of FoG and cueing activation. In questionnaire survey, 88% of the PD participants confirmed that this wearable system could effectively enhance walking, 81% thought that the system was comfortable and convenient, and 70% overcame the FoG. Therefore, the IWS makes it an effective, powerful, and convenient tool for enhancing the mobility of people with PD.

Estimation of gait parameters using leg velocity for amputee population

Quantification of key gait parameters plays an important role in assessing gait deficits in clinical research. Gait parameter estimation using lower-limb kinematics (mainly leg velocity data) has shown promise but lacks validation for the amputee population. The aim of this study is to assess the accuracy of lower-leg angular velocity to predict key gait events (toe-off and heel strike) and associated temporal parameters for the amputee population. An open data set of reflexive markers during treadmill walking from 10 subjects with unilateral transfemoral amputation was used. A rule-based dual-minima algorithm was developed to detect the landmarks in the shank velocity signal indicating toe-off and heel strike events. Four temporal gait parameters were also estimated (step time, stride time, stance and swing duration). These predictions were compared against the force platform data for 3000 walking cycles from 239 walking trials. Considerable accuracy was achieved for the HS event as well as for step and stride timings, with mean errors ranging from 0 to -13ms. The TO prediction exhibited a larger error with its mean ranging from 35-81ms. The algorithm consistently predicted the TO earlier than the actual event, resulting in prediction errors in stance and swing timings. Significant differences were found between the prediction for sound and prosthetic legs, with better TO accuracy on the prosthetic side. The prediction accuracy also appeared to improve with the subjects’ mobility level (K-level). In conclusion, the leg velocity profile, coupled with the dual-minima algorithm, can predict temporal parameters for the transfemoral amputee population with varying degrees of accuracy.

Effect of Torso Kinematics on Gait Phase Estimation at Different Walking Speeds

Human gait phase estimation has been studied in the field of robotics due to its importance for controlling wearable devices (e.g., prostheses or exoskeletons) in a synchronized manner with the user. As data-driven approaches have recently risen in the field, researchers have attempted to estimate the user gait phase using a learning-based method. Thigh and torso information have been widely utilized in estimating the human gait phase for wearable devices. Torso information, however, is known to have high variability, specifically in slow walking, and its effect on gait phase estimation has not been studied. In this study, we quantified torso variability and investigated how the torso information affects the gait phase estimation result at various walking speeds. We obtained three different trained models (i.e., general, slow, and normal-fast models) using long short-term memory (LSTM). These models were compared to identify the effect of torso information at different walking speeds. In addition, the ablation study was performed to identify the isolated effect of the torso on the gait phase estimation. As a result, when the torso segment's angular velocity was used with thigh information, the accuracy of gait phase estimation was increased, while the torso segment's angular position had no apparent effect on the accuracy. This study suggests that the torso segment's angular velocity enhances human gait phase estimation when used together with the thigh information despite its known variability.

Automatic Hemiplegia Type Detection (Right or Left) Using the Levenberg-Marquardt Backpropagation Method

Hemiplegia affects a significant portion of the human population. It is a condition that causes motor impairment and severely reduces the patient’s quality of life. This paper presents an automatic system for identifying the hemiplegia type (right or left part of the body is affected). The proposed system utilizes the data taken from patients and healthy subjects using the accelerometer sensor from the RehaGait mobile gait analysis system. The collected data undergo a pre-processing procedure followed by a feature extraction stage. The extracted features are then sent to a neural network trained by the Levenberg-Marquardt backpropagation (LM-BP) algorithm. The experimental part of this research involved creating a custom-created dataset containing entries taken from ten healthy and twenty non-healthy subjects. The data were taken from seven different sensors placed in specific areas of the subjects’ bodies. These sensors can capture a three-dimensional (3D) signal using the accelerometer, magnetometer, and gyroscope device types. The proposed system used the signals taken from the accelerometers, which were split into 2-sec windows. The proposed system achieved a classification accuracy of 95.12% and was compared with fourteen commonly used machine learning approaches.

Smartphone-based human fatigue level detection using machine learning approaches

Human muscle fatigue is the main result of diminishing muscle capability, leading to reduced performance and increased risk of falls and injury. This study provides a classification model to identify the human fatigue level based on the motion signals collected by a smartphone. 24 participants were recruited and performed the fatiguing exercise (i.e. squatting). Upon completing each set of squatting, they walked for a fixed distance while the smartphone attached to their right shank and the gait data were associated with the Borg's Rating of Perceived Exertion (i.e. data label). Our machine-learning model of two (no- vs. strong-fatigue), three (no-, medium-, and strong-fatigue) and four (no-, low-, medium-, and strong-fatigue) levels of fatigue reached the accuracy of 91, 78, and 64%, respectively. The outcomes of this study may facilitate the accessibility of a fatigue-monitoring tool in the workplace, which improves the workers' performance and reduce the risk of falls and injury. Practitioner Summary: This study aimed to develop a machine-learning model to identify human fatigue level using motion data captured by a smartphone attached to the shank. Our results can facilitate the development of an accessible fatigue-monitoring system that may improve the workers' performance and reduce the risk of falls and injury. Abbreviations: WMSD: work-related musculoskeletal disorders; IMU: inertial measurement unit; RPE: rating of perceived exertion; SVM: support vector machine; IRB: institutional review board; SOM: self-organizing map; LDA: linear discriminant analysis; PCA: principal component analysis; FT: fourier transformation; RBF: radial basis function; CUSUM: cumulative sum; ROM: range of motion; MVC: maximum voluntary contractions.

Wearable Sensor-Based Real-Time Gait Detection: A Systematic Review

Gait analysis has traditionally been carried out in a laboratory environment using expensive equipment, but, recently, reliable, affordable, and wearable sensors have enabled integration into clinical applications as well as use during activities of daily living. Real-time gait analysis is key to the development of gait rehabilitation techniques and assistive devices such as neuroprostheses. This article presents a systematic review of wearable sensors and techniques used in real-time gait analysis, and their application to pathological gait. From four major scientific databases, we identified 1262 articles of which 113 were analyzed in full-text. We found that heel strike and toe off are the most sought-after gait events. Inertial measurement units (IMU) are the most widely used wearable sensors and the shank and foot are the preferred placements. Insole pressure sensors are the most common sensors for ground-truth validation for IMU-based gait detection. Rule-based techniques relying on threshold or peak detection are the most widely used gait detection method. The heterogeneity of evaluation criteria prevented quantitative performance comparison of all methods. Although most studies predicted that the proposed methods would work on pathological gait, less than one third were validated on such data. Clinical applications of gait detection algorithms were considered, and we recommend a combination of IMU and rule-based methods as an optimal solution.

Identification of Patients with Sarcopenia Using Gait Parameters Based on Inertial Sensors

Sarcopenia can cause various senile diseases and is a major factor associated with the quality of life in old age. To diagnose, assess, and monitor muscle loss in daily life, 10 sarcopenia and 10 normal subjects were selected using lean mass index and grip strength, and their gait signals obtained from inertial sensor-based gait devices were analyzed. Given that the inertial sensor can measure the acceleration and angular velocity, it is highly useful in the kinematic analysis of walking. This study detected spatial-temporal parameters used in clinical practice and descriptive statistical parameters for all seven gait phases for detailed analyses. To increase the accuracy of sarcopenia identification, we used Shapley Additive explanations to select important parameters that facilitated high classification accuracy. Support vector machines (SVM), random forest, and multilayer perceptron are classification methods that require traditional feature extraction, whereas deep learning methods use raw data as input to identify sarcopenia. As a result, the input that used the descriptive statistical parameters for the seven gait phases obtained higher accuracy. The knowledge-based gait parameter detection was more accurate in identifying sarcopenia than automatic feature selection using deep learning. The highest accuracy of 95% was achieved using an SVM model with 20 descriptive statistical parameters. Our results indicate that sarcopenia can be monitored with a wearable device in daily life.

Identification of Gait Events in Healthy Subjects and With Parkinson's Disease Using Inertial Sensors: An Adaptive Unsupervised Learning Approach

Automatic identification of gait events is an essential component of the control scheme of assistive robotic devices. Many available techniques suffer limitations for real-time implementations and in guaranteeing high performances when identifying events in subjects with gait impairments. Machine learning algorithms offer a solution by enabling the training of different models to represent the gait patterns of different subjects. Here our aim is twofold: to remove the need for training stages using unsupervised learning, and to modify the parameters according to the changes within a walking trial using adaptive procedures. We developed two adaptive unsupervised algorithms for real-time detection of four gait events, using only signals from two single-IMU foot-mounted wearable devices. We evaluated the algorithms using data collected from five healthy adults and seven subjects with Parkinson's disease (PD) walking overground and on a treadmill. Both algorithms obtained high performance in terms of accuracy ( F₁ -score ≥ 0.95 for both groups), and timing agreement using a force-sensitive resistors as reference (mean absolute differences of 66 ± 53 msec for the healthy group, and 58 ± 63 msec for the PD group). The proposed algorithms demonstrated the potential to learn optimal parameters for a particular participant and for detecting gait events without additional sensors, external labeling, or long training stages.

Finite Class Bayesian Inference System for Circle and Linear Walking Gait Event Recognition Using Inertial Measurement Units

Accurate and fast human motion pattern recognition is the key to ensuring lower limb assistive devices' appropriate assistance. The research on human motion pattern recognition of lower limb assistive devices mainly focuses on sagittal gait. The motion pattern such as circular walking (CW) is asymmetric about the sagittal plane of the body. CW is common in daily living. However, the recognition algorithm of CW is rarely reported. Since lower limb assistive devices interact with humans, lacking the capability of recognizing CW is dangerous. Thus, to realize the accurate and fast recognition of CW, this article proposed a finite class Bayesian interference system (FC-BesIS). FC-BesIS is designed to recognize walking activities (linear walking and CW) and gait events (heel contact, load response, mid stance, terminal stance, pre-swing, initial swing, mid swing, and terminal swing). A finite class method which reduces the number of potential classes according to elimination rules before decision-making is introduced. Elimination rules are designed based on likelihood estimation and sensor information. The experiments show that walking activities and gait events can be accurately and fastly recognized by FC-BesIS. The experiments also show that the performance of FC-BesIS in mean recognition accuracy (MRA) and mean decision time (MDT) is improved compared with BesIS. The MRA of walking activities and gait events are 100% and 97.38%, respectively. The MDT of walking activities and gait events are 28.19 ms and 33.94 ms, respectively. Overall, FC-BesIS has been proved to be an accurate and fast recognition algorithm for human motion patterns using wearable sensors.

A review of foot pose and trajectory estimation methods using inertial and auxiliary sensors for kinematic gait analysis

The use of foot mounted inertial and other auxiliary sensors for kinematic gait analysis has been extensively investigated during the last years. Although, these sensors still yield less accurate results than those obtained employing optical motion capture systems, the miniaturization and their low cost have allowed the estimation of kinematic spatiotemporal parameters in laboratory conditions and real life scenarios. The aim of this work was to present a comprehensive approach of this scientific area through a systematic literature research, breaking down the state-of-the-art methods into three main parts: (1) zero velocity interval detection techniques; (2) assumptions and sensors' utilization; (3) foot pose and trajectory estimation methods. Published articles from 1995 until December of 2018 were searched in the PubMed, IEEE Xplore and Google Scholar databases. The research was focused on two categories: (a) zero velocity interval detection methods; and (b) foot pose and trajectory estimation methods. The employed assumptions and the potential use of the sensors have been identified from the retrieved articles. Technical characteristics, categorized methodologies, application conditions, advantages and disadvantages have been provided, while, for the first time, assumptions and sensors' utilization have been identified, categorized and are presented in this review. Considerable progress has been achieved in gait parameters estimation on constrained laboratory environments taking into account assumptions such as a person walking on a flat floor. On the contrary, methods that rely on less constraining assumptions, and are thus applicable in daily life, led to less accurate results. Rule based methods have been mainly used for the detection of the zero velocity intervals, while more complex techniques have been proposed, which may lead to more accurate gait parameters. The review process has shown that presently the best-performing methods for gait parameter estimation make use of inertial sensors combined with auxiliary sensors such as ultrasonic sensors, proximity sensors and cameras. However, the experimental evaluation protocol was much more thorough, when single inertial sensors were used. Finally, it has been highlighted that the accuracy of setups using auxiliary sensors may further be improved by collecting measurements during the whole foot movement and not only partially as is currently the practice. This review has identified the need for research and development of methods and setups that allow for the robust estimation of kinematic gait parameters in unconstrained environments and under various gait profiles.

Gait Cycle Validation and Segmentation Using Inertial Sensors

In this paper, we develop an algorithm to automatically validate and segment a gait cycle in real time into three gait events, namely midstance, toe-off, and heel-strike, using inertial sensors. We first use the physical models of sensor data obtained from a foot-mounted inertial system to differentiate stationary and moving segments of the sensor data. Next, we develop an optimization routine called sparsity-assisted wavelet denoising (SAWD), which simultaneously combines linear time invariant filters, orthogonal multiresolution representations such as wavelets, and sparsity-based methods, to generate a sparse template of the moving segments of the gyroscope measurements in the sagittal plane for valid gait cycles. Thereafter, to validate any moving segment as a gait cycle, we compute the root-mean-square error between the generated sparse template and the sparse representation of the moving segment of the gyroscope data in the sagittal plane obtained using SAWD. Finally, we find the local minima for the stationary and moving segments of a valid gait cycle to detect the gait events. We compare our proposed method with existing methods, for a fixed threshold, using real data obtained from three groups, namely controls, participants with Parkinson disease, and geriatric participants. Our proposed method demonstrates an average F1 score of 87.78% across all groups for a fixed sampling rate, and an average F1 score of 92.44% across all Parkinson disease participants for a variable sampling rate.

Gait Event Detection in Controlled and Real-Life Situations: Repeated Measures From Healthy Subjects

A benchmark and time-effective computational method is needed to assess human gait events in real-life walking situations using few sensors to be easily reproducible. This paper fosters a reliable gait event detection system that can operate at diverse gait speeds and on diverse real-life terrains by detecting several gait events in real time. This detection only relies on the foot angular velocity measured by a wearable gyroscope mounted in the foot to facilitate its integration for daily and repeated use. To operate as a benchmark tool, the proposed detection system endows an adaptive computational method by applying a finite-state machine based on heuristic decision rules dependent on adaptive thresholds. Repeated measurements from 11 healthy subjects (28.27 ± 4.17 years) were acquired in controlled situations through a treadmill at different speeds (from 1.5 to 4.5 km/h) and slopes (from 0% to 10%). This validation also includes heterogeneous gait patterns from nine healthy subjects (27 ± 7.35 years) monitored at three self-selected paces (from 1 ± 0.2 to 2 ± 0.18 m/s) during forward walking on flat, rough, and inclined surfaces and climbing staircases. The proposed method was significantly more accurate ( ) and time effective (< 30.53 ± 9.88 ms, ) in a benchmarking analysis with a state-of-the-art method during 5657 steps. Heel strike was the gait event most accurately detected under controlled (accuracy of 100%) and real-life situations (accuracy > 96.98%). Misdetection was more pronounced in middle mid swing (accuracy > 90.12%). The lower computational load, together with an improved performance, makes this detection system suitable for quantitative benchmarking in the locomotor rehabilitation field.

POF Smart Carpet: A Multiplexed Polymer Optical Fiber-Embedded Smart Carpet for Gait Analysis

This paper presents the development of a smart carpet based on polymer optical fiber (POF) for ground reaction force (GRF) and spatio-temporal gait parameter assessment. The proposed carpet has 20 intensity variation-based sensors on one fiber with two photodetectors for acquisition, each one for the response of 10 closer sensors. The used multiplexing technique is based on side-coupling between the light sources and POF lateral sections in which one light-emitting diode (LED) is activated at a time, sequentially. Three tests were performed, two for sensor characterization and one for validation of the smart carpet, where the first test consisted of the application of calibrated weights on the top of each sensor for force characterization. In the second test, the foot was positioned on predefined points distributed on the carpet, where a mean relative error of 2.9% was obtained. Results of the walking tests on the proposed POF-embedded smart carpet showed the possibility of estimating the GRF and spatio-temporal gait parameters (step and stride lengths, cadence, and stance duration). The obtained results make possible the identification of gait events (stance and swing phases) as well as the stance duration and double support periods. The proposed carpet is a low-cost and reliable tool for gait analysis in different applications.

Gait Phase Detection for Lower-Limb Exoskeletons using Foot Motion Data from a Single Inertial Measurement Unit in Hemiparetic Individuals

Due to the recent rise in the use of lower-limb exoskeletons as an alternative for gait rehabilitation, gait phase detection has become an increasingly important feature in the control of these devices. In addition, highly functional, low-cost recovery devices are needed in developing countries, since limited budgets are allocated specifically for biomedical advances. To achieve this goal, this paper presents two gait phase partitioning algorithms that use motion data from a single inertial measurement unit (IMU) placed on the foot instep. For these data, sagittal angular velocity and linear acceleration signals were extracted from nine healthy subjects and nine pathological subjects. Pressure patterns from force sensitive resistors (FSR) instrumented on a custom insole were used as reference values. The performance of a threshold-based (TB) algorithm and a hidden Markov model (HMM) based algorithm, trained by means of subject-specific and standardized parameters approaches, were compared during treadmill walking tasks in terms of timing errors and the goodness index. The findings indicate that HMM outperforms TB for this hardware configuration. In addition, the HMM-based classifier trained by an intra-subject approach showed excellent reliability for the evaluation of mean time, i.e., its intra-class correlation coefficient (ICC) was greater than 0 . 75 . In conclusion, the HMM-based method proposed here can be implemented for gait phase recognition, such as to evaluate gait variability in patients and to control robotic orthoses for lower-limb rehabilitation.

Automatic Detection of Faults in Race Walking: A Comparative Analysis of Machine-Learning Algorithms Fed with Inertial Sensor Data

The validity of results in race walking is often questioned due to subjective decisions in the detection of faults. This study aims to compare machine-learning algorithms fed with data gathered from inertial sensors placed on lower-limb segments to define the best-performing classifiers for the automatic detection of illegal steps. Eight race walkers were enrolled and linear accelerations and angular velocities related to pelvis, thighs, shanks, and feet were acquired by seven inertial sensors. The experimental protocol consisted of two repetitions of three laps of 250 m, one performed with regular race walking, one with loss-of-contact faults, and one with knee-bent faults. The performance of 108 classifiers was evaluated in terms of accuracy, recall, precision, F1-score, and goodness index. Generally, linear accelerations revealed themselves as more characteristic with respect to the angular velocities. Among classifiers, those based on the support vector machine (SVM) were the most accurate. In particular, the quadratic SVM fed with shank linear accelerations was the best-performing classifier, with an F1-score and a goodness index equal to 0.89 and 0.11, respectively. The results open the possibility of using a wearable device for automatic detection of faults in race walking competition.

Using a deep learning network to recognise low back pain in static standing

Low back pain (LBP) remains one of the most prevalent musculoskeletal disorders, while algorithms that able to recognise LBP patients from healthy population using balance performance data are rarely seen. In this study, human balance and body sway performance during standing trials were utilised to recognise chronic LBP populations using deep neural networks. To be specific, 44 chronic LBP and healthy individuals performed static standing tasks, while their spine kinematics and centre of pressure were recorded. A deep learning network with long short-term memory units was used for training, prediction and implementation. The performance of the model was evaluated by: (a) overall accuracy, (b) precision, (c) recall, (d) F1 measure, (e) receiver-operating characteristic and (f) area under the curve. Results indicated that deep neural networks could recognise LBP populations with precision up to 97.2% and recall up to 97.2%. Meanwhile, the results showed that the model with the C7 sensor output performed the best. Practitioner summary: Low back pain (LBP) remains the most common musculoskeletal disorder. In this study, we investigated the feasibility of applying artificial intelligent deep neural network in detecting LBP population from healthy controls with their kinematics data. Results showed a deep learning network can solve the above classification problem with both promising precision and recall performance.

Longitudinal Walking Analysis in Hemiparetic Patients Using Wearable Motion Sensors: Is There Convergence Between Body Sides?

Background: Longitudinal movement parameter analysis of hemiparetic patients over several months could reveal potential recovery trends and help clinicians adapting therapy strategies to maximize recovery outcome. Wearable sensors offer potential for day-long movement recordings in realistic rehabilitation settings including activities of daily living, e.g., walking. The measurement of walking-related movement parameters of affected and non-affected body sides are of interest to determine mobility and investigate recovery trends. Methods: By comparing movement of both body sides, recovery trends across the rehabilitation duration were investigated. We derived and validated selected walking segments from free-living, day-long movement by using rules that do not require data-based training or data annotations. Automatic stride segmentation using peak detection was applied to walking segments. Movement parameters during walking were extracted, including stride count, stride duration, cadence, and sway. Finally, linear regression models over each movement parameter were derived to forecast the moment of convergence between body sides. Convergence points were expressed as duration and investigated in a patient observation study. Results: Convergence was analyzed in walking-related movement parameters in an outpatient study including totally 102 full-day recordings of inertial movement data from 11 hemiparetic patients. The recordings were performed over several months in a day-care centre. Validation of the walking extraction method from sensor data yielded sensitivities up to 80 % and specificity above 94 % on average. Comparison of automatically and manually derived movement parameters showed average relative errors below 6 % between affected and non-affected body sides. Movement parameter variability within and across patients was observed and confirmed by case reports, reflecting individual patient behavior. Conclusion: Convergence points were proposed as intuitive metric, which could facilitate training personalization for patients according to their individual needs. Our continuous movement parameter extraction and analysis, was feasible for realistic, day-long recordings without annotations. Visualizations of movement parameter trends and convergence points indicated that individual habits and patient therapies were reflected in walking and mobility. Context information of clinical case reports supported trend and convergence interpretation. Inconsistent convergence point estimation suggested individually varying deficiencies. Long-term recovery monitoring using convergence points could support patient-specific training strategies in future remote rehabilitation.

On the security of consumer wearable devices in the Internet of Things

Miniaturization of computer hardware and the demand for network capable devices has resulted in the emergence of a new class of technology called wearable computing. Wearable devices have many purposes like lifestyle support, health monitoring, fitness monitoring, entertainment, industrial uses, and gaming. Wearable devices are hurriedly being marketed in an attempt to capture an emerging market. Owing to this, some devices do not adequately address the need for security. To enable virtualization and connectivity wearable devices sense and transmit data, therefore it is essential that the device, its data and the user are protected. In this paper the use of novel Integrated Circuit Metric (ICMetric) technology for the provision of security in wearable devices has been suggested. ICMetric technology uses the features of a device to generate an identification which is then used for the provision of cryptographic services. This paper explores how a device ICMetric can be generated by using the accelerometer and gyroscope sensor. Since wearable devices often operate in a group setting the work also focuses on generating a group identification which is then used to deliver services like authentication, confidentiality, secure admission and symmetric key generation. Experiment and simulation results prove that the scheme offers high levels of security without compromising on resource demands.

In Memoriam: Paolo Cappa

Prof. Paolo Cappa passed away on 26 August 2016, at the age of 59, after a long and courageous fight against cancer. Paolo Cappa was a Professor in Mechanical and Thermal Measurements and Experimental Biomechanics in the Department of Mechanical and Aerospace Engineering of Sapienza University of Rome, where he had also served as the Head of the Department, and a Research Professor in the Department of Mechanical and Aerospace Engineering of New York University Tandon School of Engineering. During his intense, yet short, career, he made several significant scientific contributions within the discipline of Mechanical and Thermal Measurements, pioneering fundamental applications to Biomechanics. He co-founded the Motion Analysis and Robotics Laboratory (MARLab) within the Neurorehabilitation Division of IRCCS Pediatric Hospital “Bambino Gesu”, in Rome, to fuel transitional research from the laboratory to clinical practice. Through collaboration with neurologists and physiatrists at MARLab, Prof. Cappa led the development of a powerful array of novel mechanical solutions to wearable robotics for pediatric patients, addressing dramatic needs for children’s health and contributing to the training of an entire generation of Mechanical Engineering students.

An Acceleration-Based Gait Assessment Method for Children with Cerebral Palsy

With the aim of providing an objective tool for motion disability assessment in clinical diagnosis and rehabilitation therapy of cerebral palsy (CP) patients, an acceleration-based gait assessment method was proposed in this paper. To capture gait information, three inertial measurement units (IMUs) were placed on the lower trunk and thigh, respectively. By comparing differences in the gait acceleration modes between children with CP and healthy subjects, an assessment method based on grey relational analysis and five gait parameters, including Pearson coefficient, variance ratio, the number of extreme points, harmonic ratio and symmetry was established. Twenty-two children with cerebral palsy (7.49 ± 2.86 years old), fourteen healthy adults (24.2 ± 1.55 years old) and ten healthy children (7.03 ± 1.49 years old) participated in the gait data acquisition experiment. The results demonstrated that, compared to healthy subjects, the symptoms and severity of motor dysfunction of CP children could result in abnormality of the gait acceleration modes, and the proposed assessment method was able to effectively evaluate the degree gait abnormality in CP children.

Automatic classification of gait in children with early-onset ataxia or developmental coordination disorder and controls using inertial sensors

Early-Onset Ataxia (EOA) and Developmental Coordination Disorder (DCD) are two conditions that affect coordination in children. Phenotypic identification of impaired coordination plays an important role in their diagnosis. Gait is one of the tests included in rating scales that can be used to assess motor coordination. A practical problem is that the resemblance between EOA and DCD symptoms can hamper their diagnosis. In this study we employed inertial sensors and a supervised classifier to obtain an automatic classification of the condition of participants. Data from shank and waist mounted inertial measurement units were used to extract features during gait in children diagnosed with EOA or DCD and age-matched controls. We defined a set of features from the recorded signals and we obtained the optimal features for classification using a backward sequential approach. We correctly classified 80.0%, 85.7%, and 70.0% of the control, DCD and EOA children, respectively. Overall, the automatic classifier correctly classified 78.4% of the participants, which is slightly better than the phenotypic assessment of gait by two pediatric neurologists (73.0%). These results demonstrate that automatic classification employing signals from inertial sensors obtained during gait maybe used as a support tool in the differential diagnosis of EOA and DCD. Furthermore, future extension of the classifier's test domains may help to further improve the diagnostic accuracy of pediatric coordination impairment. In this sense, this study may provide a first step towards incorporating a clinically objective and viable biomarker for identification of EOA and DCD.

Lessons Learned from Implementing a Rapid Test of a Technology Device in a Tertiary Hospital in Uganda

BACKGROUND

Many African hospitals participate in technology research studies that take many months or years. Fewer sites have experience with rapid studies, conducted over a period of weeks. Such studies can benefit the institution and its patients in the short term, and in the long term can help prepare the institution for adopting the new technology.

OBJECTIVES

We conducted a rapid validation study of consumer fitness device at Mulago National Referral and Teaching Hospital in Kampala, Uganda. In doing so, we captured valuable lessons about how to conduct a rapid study that will be useful to future researchers conducting similar fast-paced studies.

METHODS

We conducted a descriptive study of a convenience sample of 57 patients. Patients who volunteered wore a fitness wristband device. Study staff collected vital signs using standard approaches.

FINDINGS

Our findings were as follows: (1) effective partnership by local experts can ensure success; (2) a PI with experience working with the hospital ethics committee is essential to a rapid study; (3) reassurance that the study design benefits patients and the institution can help speed approval; (4) conduct detailed assessment of patient population in advance; (5) allow sufficient time for logistics arrangements; (6) quickly pivot the approach as needed, consistent with the protocol; (7) conduct data quality review on every shift; (8) conduct a supplies inventory at the end of each shift; (9) make rapid decisions about hiring and discontinuing study staff; (10) implement a patient location protocol at the start of the study; and (11) ensure availability of study staff refreshments in the study room.

CONCLUSION

A rapid study of innovative technology can be successful at a hospital in a resource-limited setting.

Gait Phase Recognition for Lower-Limb Exoskeleton with Only Joint Angular Sensors

Gait phase is widely used for gait trajectory generation, gait control and gait evaluation on lower-limb exoskeletons. So far, a variety of methods have been developed to identify the gait phase for lower-limb exoskeletons. Angular sensors on lower-limb exoskeletons are essential for joint closed-loop controlling; however, other types of sensors, such as plantar pressure, attitude or inertial measurement unit, are not indispensable.Therefore, to make full use of existing sensors, we propose a novel gait phase recognition method for lower-limb exoskeletons using only joint angular sensors. The method consists of two procedures. Firstly, the gait deviation distances during walking are calculated and classified by Fisher’s linear discriminant method, and one gait cycle is divided into eight gait phases. The validity of the classification results is also verified based on large gait samples. Secondly, we build a gait phase recognition model based on multilayer perceptron and train it with the phase-labeled gait data. The experimental result of cross-validation shows that the model has a 94.45% average correct rate of set (CRS) and an 87.22% average correct rate of phase (CRP) on the testing set, and it can predict the gait phase accurately. The novel method avoids installing additional sensors on the exoskeleton or human body and simplifies the sensory system of the lower-limb exoskeleton.

A Machine Learning Framework for Gait Classification Using Inertial Sensors: Application to Elderly, Post-Stroke and Huntington's Disease Patients

Machine learning methods have been widely used for gait assessment through the estimation of spatio-temporal parameters. As a further step, the objective of this work is to propose and validate a general probabilistic modeling approach for the classification of different pathological gaits. Specifically, the presented methodology was tested on gait data recorded on two pathological populations (Huntington’s disease and post-stroke subjects) and healthy elderly controls using data from inertial measurement units placed at shank and waist. By extracting features from group-specific Hidden Markov Models (HMMs) and signal information in time and frequency domain, a Support Vector Machines classifier (SVM) was designed and validated. The 90.5% of subjects was assigned to the right group after leave-one-subject–out cross validation and majority voting. The long-term goal we point to is the gait assessment in everyday life to early detect gait alterations.

Gait Partitioning Methods: A Systematic Review

In the last years, gait phase partitioning has come to be a challenging research topic due to its impact on several applications related to gait technologies. A variety of sensors can be used to feed algorithms for gait phase partitioning, mainly classifiable as wearable or non-wearable. Among wearable sensors, footswitches or foot pressure insoles are generally considered as the gold standard; however, to overcome some inherent limitations of the former, inertial measurement units have become popular in recent decades. Valuable results have been achieved also though electromyography, electroneurography, and ultrasonic sensors. Non-wearable sensors, such as opto-electronic systems along with force platforms, remain the most accurate system to perform gait analysis in an indoor environment. In the present paper we identify, select, and categorize the available methodologies for gait phase detection, analyzing advantages and disadvantages of each solution. Finally, we comparatively examine the obtainable gait phase granularities, the usable computational methodologies and the optimal sensor placements on the targeted body segments.

Validation of Inter-Subject Training for Hidden Markov Models Applied to Gait Phase Detection in Children with Cerebral Palsy

Gait-phase recognition is a necessary functionality to drive robotic rehabilitation devices for lower limbs. Hidden Markov Models (HMMs) represent a viable solution, but they need subject-specific training, making data processing very time-consuming. Here, we validated an inter-subject procedure to avoid the intra-subject one in two, four and six gait-phase models in pediatric subjects. The inter-subject procedure consists in the identification of a standardized parameter set to adapt the model to measurements. We tested the inter-subject procedure both on scalar and distributed classifiers. Ten healthy children and ten hemiplegic children, each equipped with two Inertial Measurement Units placed on shank and foot, were recruited. The sagittal component of angular velocity was recorded by gyroscopes while subjects performed four walking trials on a treadmill. The goodness of classifiers was evaluated with the Receiver Operating Characteristic. The results provided a goodness from good to optimum for all examined classifiers (0 < G < 0.6), with the best performance for the distributed classifier in two-phase recognition (G = 0.02). Differences were found among gait partitioning models, while no differences were found between training procedures with the exception of the shank classifier. Our results raise the possibility of avoiding subject-specific training in HMM for gait-phase recognition and its implementation to control exoskeletons for the pediatric population.

Solutions to the Inverse LQR Problem with Application to Biological Systems Analysis

In this paper, we present a set of techniques for finding a cost function to the time-invariant Linear Quadratic Regulator (LQR) problem in both continuous- and discrete-time cases. Our methodology is based on the solution to the inverse LQR problem, which can be stated as: does a given controller K describe the solution to a time-invariant LQR problem, and if so, what weights Q and R produce K as the optimal solution? Our motivation for investigating this problem is the analysis of motion goals in biological systems. We first describe an efficient Linear Matrix Inequality (LMI) method for determining a solution to the general case of this inverse LQR problem when both the weighting matrices Q and R are unknown. Our first LMI-based formulation provides a unique solution when it is feasible. Additionally, we propose a gradient-based, least-squares minimization method that can be applied to approximate a solution in cases when the LMIs are infeasible. This new method is very useful in practice since the estimated gain matrix K from the noisy experimental data could be perturbed by the estimation error, which may result in the infeasibility of the LMIs. We also provide an LMI minimization problem to find a good initial point for the minimization using the proposed gradient descent algorithm. We then provide a set of examples to illustrate how to apply our approaches to several different types of problems. An important result is the application of the technique to human subject posture control when seated on a moving robot. Results show that we can recover a cost function which may provide a useful insight on the human motor control goal.

A novel HMM distributed classifier for the detection of gait phases by means of a wearable inertial sensor network

In this work, we decided to apply a hierarchical weighted decision, proposed and used in other research fields, for the recognition of gait phases. The developed and validated novel distributed classifier is based on hierarchical weighted decision from outputs of scalar Hidden Markov Models (HMM) applied to angular velocities of foot, shank, and thigh. The angular velocities of ten healthy subjects were acquired via three uni-axial gyroscopes embedded in inertial measurement units (IMUs) during one walking task, repeated three times, on a treadmill. After validating the novel distributed classifier and scalar and vectorial classifiers-already proposed in the literature, with a cross-validation, classifiers were compared for sensitivity, specificity, and computational load for all combinations of the three targeted anatomical segments. Moreover, the performance of the novel distributed classifier in the estimation of gait variability in terms of mean time and coefficient of variation was evaluated. The highest values of specificity and sensitivity (>0.98) for the three classifiers examined here were obtained when the angular velocity of the foot was processed. Distributed and vectorial classifiers reached acceptable values (>0.95) when the angular velocity of shank and thigh were analyzed. Distributed and scalar classifiers showed values of computational load about 100 times lower than the one obtained with the vectorial classifier. In addition, distributed classifiers showed an excellent reliability for the evaluation of mean time and a good/excellent reliability for the coefficient of variation. In conclusion, due to the better performance and the small value of computational load, the here proposed novel distributed classifier can be implemented in the real-time application of gait phases recognition, such as to evaluate gait variability in patients or to control active orthoses for the recovery of mobility of lower limb joints.

Experimental evaluation of indoor magnetic distortion effects on gait analysis performed with wearable inertial sensors

Magnetic inertial measurement unit systems (MIMU) offer the potential to perform joint kinematics evaluation as an alternative to optoelectronic systems (OS). Several studies have reported the effect of indoor magnetic field disturbances on the MIMU's heading output, even though the overall effect on the evaluation of lower limb joint kinematics is not yet fully explored. The aim of the study is to assess the influence of indoor magnetic field distortion on gait analysis trials conducted with a commercial MIMU system. A healthy adult performed gait analysis sessions both indoors and outdoors. Data collected indoors were post-processed with and without a heading correction methodology performed with OS at the start of the gait trial. The performance of the MIMU system is characterized in terms of indices, based on the mean value of lower limb joint angles and the associated ROM, quantifying the system repeatability. We find that the effects of magnetic field distortion, such as the one we experienced in our lab, were limited to the transverse plane of each joint and to the frontal plane of the ankle. Sagittal plane values, instead, showed sufficient repeatability moving from outdoors to indoors. Our findings provide indications to clinicians on MIMU performance in the measurement of lower limb kinematics.