Inertial Sensor Measurements Research Articles

The Short Physical Performance Battery does not correlate with daily life gait quality and quantity in community-dwelling older adults with an increased fall risk

BackgroundBoth the Short Physical Performance Battery (SPPB) and daily life gait quality and quantity obtained from wearable sensors are used to measure functional status in older adults. It is generally assumed that they are interrelated and exchangeable, but this has not yet been established. Interchangeability of these measures would pave the way for remote monitoring of functional status. Research QuestionAre the SPPB and daily life gait quality and quantity measures correlated in community-dwelling older adults? MethodsThe SPPB and gait quality and quantity data of 229 community-dwelling adults of 65 years or older were collected. The SPPB is a combined score of the Three Stage Balance test, Four Meter Walk test, and Five Times Sit to Stand test and ranges from 0 to 12. Participants wore a tri-axial inertial sensor for one week to assess gait quality (e.g. gait stability and smoothness) and quantity (e.g. number of strides). Correlation coefficients between SPPB scores and gait quality and quantity measures were assessed using Spearman’s correlation. ResultsThe median age of the study population was 76.2 years (IQR 72.6–81.0), and 76 % were women (n=175). The median SPPB score was 10 (IQR 8–11). Spearman's correlation coefficients between the SPPB and gait quality and quantity measures were all below 0.3. SignificanceA possible explanation for the observed weak correlations is that the SPPB reflects one’s maximal capacity, while gait quality and quantity reflect the submaximal performance in daily life. The SPPB and gait quality and quantity seem therefore distinct constructs with complementary value, rather than interchangeable. A more comprehensive understanding of functional status might be achieved by combining the SPPB assessment of standardized activities with the evaluation of inertial sensor measurements obtained during daily life activities.

Serial kinematic analysis using inertial measurement units in growing dogs at risk of hip dysplasia

In this study, 54 dogs were examined at regular intervals from 12 weeks to 15 months of age using a gait analysis system based on inertial measurement sensors. At the end of the study, the dogs were examined for hip dysplasia (HD) and elbow dysplasia (ED) under sedation and officially classified at the Dysplasia Commission in Zurich. Gait parameters which are characteristic for the gait pattern of dogs, were calculated according to recent publications. These parameters were analysed for variance throughout the entire study period and assigned to healthy dogs and those suffering from HD. The findings of the study show that dogs suffering from HD exhibit a more unsteady gait pattern, e.g. higher variance, as they grow.

Exposures to select risk factors can be estimated from a continuous stream of inertial sensor measurements during a variety of lifting-lowering tasks

Wearable inertial measurement units (IMUs) are used increasingly to estimate biomechanical exposures in lifting-lowering tasks. The objective of the study was to develop and evaluate predictive models for estimating relative hand loads and two other critical biomechanical exposures to gain a comprehensive understanding of work-related musculoskeletal disorders in lifting. We collected 12,480 lifting-lowering phases from 26 subjects (15 men and 11 women) performing manual lifting-lowering tasks with hand loads (0–22.7 kg) at varied workstation heights and handling modes. We implemented a Hierarchical model, that sequentially classified risk factors, including workstation height, handling mode, and relative hand load. Our algorithm detected lifting-lowering phases (>97.8%) with mean onset errors of 0.12 and 0.2 seconds for lifting and lowering phases. It estimated workstation height (>98.5%), handling mode (>87.1%), and relative hand load (mean absolute errors of 5.6–5.8%) across conditions, highlighting the benefits of data-driven models in deriving lifting-lowering occurrences, timing, and critical risk factors from continuous IMU-based kinematics.

Precision Gait Stability Analysis Combining Hardware and Software

Maintaining stable gait and avoiding falls require effective control of the body’s center of mass (CoM) relative to the base of support (BoS), defined by the feet's contact points with the floor. Quadrupeds have an advantage over bipeds but share common spinal neural control mechanisms. Human walking poses stability challenges since the CoM often extends beyond the BoS. This project aims to monitor gait stability by integrating advanced hardware and software solutions. Machine learning algorithms, such as Random Forest, will be employed to predict gait instability by capturing foot positions and sensor data. Inertial measurement sensors will track stability changes, and hardware outputs will be validated against software predictions for accuracy. A mobile app developed using React Native will display predicted parameters and indicate gait instability (yes/no). In case of detected instability, a wearable buzzer will alert the user. This system promises a more accurate solution for monitoring gait stability compared to existing models, by combining real-time data collection and sophisticated algorithms

VIO-DualProNet: Visual-inertial odometry with learning based process noise covariance

Visual-inertial odometry (VIO) is a vital technique used in robotics, augmented reality, and autonomous vehicles. It combines visual and inertial measurements to accurately estimate position and orientation. Existing VIO methods assume a fixed noise covariance for the inertial uncertainty. However, accurately determining in real-time the noise variance of the inertial sensors presents a significant challenge as the uncertainty changes throughout the operation leading to suboptimal performance and reduced accuracy. To circumvent this, we propose VIO-DualProNet, a novel approach that utilizes deep learning methods to dynamically estimate the inertial measurement unit (IMU) noise uncertainty in real-time. By designing and training a deep neural network to predict inertial noise uncertainty using only inertial sensor measurements, and integrating it into the VINS-Mono algorithm, we demonstrate a substantial improvement in accuracy and robustness. Our system outperformed constant covariance methods in 9 out of 11 test sequences, with an average improvement of 25% compared to the baseline and a 12.5% improvement over the best constant covariance combination.

Vehicle Surroundings Perception Using Micro‐electromechanical Systems Inertial Sensors

Sensors enable vehicles to perceive their surroundings even while parked. Camera‐based systems for vehicle surroundings perception are already available as of today. However, camera‐based systems result in excessive power consumption and data collection. In this article, a study on always‐on surroundings perception of parked vehicles is done using micro‐electromechanical systems (MEMS) inertial sensors as an alternative to camera‐based systems. The measurements of the MEMS inertial sensors are used to detect and classify 17 different types of events. These events include accidents, vandalism, rain, as well as normal use. A convolutional neural network (CNN) is used to classify the events. The data used to train the CNN is recorded using a test vehicle in a laboratory setting and on a test track. The CNN is quantized for deployment on a CNN hardware accelerator which is packaged with the sensor as a system‐in‐package (SiP). The novel sensor system achieves a classification accuracy of 88% at a total system power dissipation of 27 μW.

Unicompartmental Knee Arthroplasty for Osteoarthritis Eliminates Lateral Thrust: Associations between Lateral Thrust Detected by Inertial Measurement Units and Clinical Outcomes.

The purpose of this study was to investigate the relationship between clinical outcomes and lateral thrust before and after unicompartmental knee arthroplasty (UKA) using inertial measurement sensor units. Eleven knees were evaluated with gait analysis. The varus angular velocity was used to evaluate lateral thrust. The femorotibial angle (FTA) and hip-knee-ankle angle (HKA) were used to evaluate lower-limb alignment, and the Oxford Knee Score (OKS) and Japanese Orthopaedic Association Score (JOA) were used to evaluate clinical outcomes. The mean pre-UKA peak varus velocity was 37.1 ± 9.8°/s, and that for post-UKA was 28.8 ± 9.1°/s (p = 0.00003), such that instabilities clearly improved. Assuming the definition of lateral thrust is when the varus angular velocity is more than 28.1°/s, 81.8% of patients had lateral thrust preoperatively, but this decreased to 55.6% postoperatively, such that the symptoms and objective findings improved. Both OKS and JOA improved after surgery. In addition, HKA was -7.9° preoperatively and -5.8° postoperatively (p = 0.024), and FTA was 181.4° preoperatively and 178.4° postoperatively (p = 0.012). There was a positive correlation between postoperative JOA and FTA, indicating that changes in postoperative alignment affected clinical outcomes. This study quantitatively evaluated the disappearance of lateral thrust by UKA, and it found that the stability can be achieved by UKA for unstable knees with lateral thrust.

High-Precision Map Construction in Degraded Long Tunnel Environments of Urban Subways

In response to the demand for high-precision point cloud mapping of subway trains in long tunnel degradation scenarios in major urban cities, we propose a map construction method based on LiDAR and inertial measurement sensors. This method comprises a tightly coupled frontend odometry system based on error Kalman filters and backend optimization using factor graphs. In the frontend odometry, inertial calculation results serve as predictions for the filter, and residuals between LiDAR points and local map plane point clouds are used for filter updates. The global pose graph is constructed based on inter-frame odometry and other constraint factors, followed by a smoothing optimization for map building. Multiple experiments in subway tunnel scenarios demonstrate that the proposed method achieves robust trajectory estimation in long tunnel scenes, where classical multi-sensor fusion methods fail due to sensor degradation. The proposed method achieves a trajectory consistency of 0.1 m in tunnel scenes, meeting the accuracy requirements for train arrival, parking, and interval operations. Additionally, in an industrial park scenario, the method is compared with ground truth provided by inertial navigation, showing an accumulated error of less than 0.2%, indicating high precision.

Outdoor activity classification using smartphone based inertial sensor measurements

Human Activity Recognition (HAR) deals with the automatic recognition of physical activities and plays a crucial role in healthcare and sports where wearable sensors and intelligent computational techniques are used. We propose a HAR algorithm that uses the smartphones accelerometer data for human activity recognition. In particular, we present a recurrent convolutional neural network-based HAR algorithm that combines a Convolutional Neural Network (CNN) to extract temporal features from the sensor data, a Fuzzy C-Means (FCM) clustering algorithm to cluster the features extracted by the CNN, and a Long Short-Term Memory (LSTM) network to learn the temporal dependencies between the features. We evaluate the proposed methodology on two distinct datasets: the MotionSense dataset and the WISDM dataset. We evaluate the proposed CNN-FCM-LSTM model on the publicly available MotionSense dataset to classify ten activity types: 1) walking upstairs, 2) walking downstairs, 3) jogging, 4) sitting, 5) standing, 6) level ground walking, 7) jumping jacks, 8) brushing teeth, 9) writing, and 10) eating. Next, we evaluate the model’s performance on the WISDM dataset to assess its ability to generalize to unseen data. On the MotionSense test dataset, CNN-FCM-LSTM achieves a classification accuracy of 99.69%, a sensitivity of 99.62%, a specificity of 99.63%, and a false positive rate per hour (FPR/h) of 0.37%. Meanwhile, it achieves a classification accuracy of 97.27% on the WISDM dataset. The CNN-FCM-LSTM model’s capability to classify a diverse range of activities within a single architecture is noteworthy. The results suggest that the proposed CNN-FCM-LSTM model using smartphone inputs is more accurate, reliable, and robust in detecting and classifying activities than the state-of-the-art models. It should be noted that activity recognition technology has the potential to aid in studying the underpinnings of physical activity, designing more effective training regimens, and simulating the rigors of competition in sports.

Digital Measures of Postural Sway Quantify Balance Deficits in Spinocerebellar Ataxia.

Maintaining balance is crucial for independence and quality of life. Loss of balance is a hallmark of spinocerebellar ataxia (SCA). The aim of this study was to identify which standing balance conditions and digital measures of body sway were most discriminative, reliable, and valid for quantifying balance in SCA. Fifty-three people with SCA (13 SCA1, 13 SCA2, 14 SCA3, and 13 SCA6) and Scale for Assessment and Rating of Ataxia (SARA) scores 9.28 ± 4.36 and 31 healthy controls were recruited. Subjects stood in six test conditions (natural stance, feet together and tandem, each with eyes open [EO] and eyes closed [EC]) with an inertial sensor on their lower back for 30 seconds (×2). We compared test completion rate, test-retest reliability, and areas under the receiver operating characteristic curve (AUC) for seven digital sway measures. Pearson's correlations related sway with the SARA and the Patient-Reported Outcome Measure of Ataxia (PROM ataxia). Most individuals with SCA (85%-100%) could stand for 30 seconds with natural stance EO or EC, and with feet together EO. The most discriminative digital sway measures (path length, range, area, and root mean square) from the two most reliable and discriminative conditions (natural stance EC and feet together EO) showed intraclass correlation coefficients from 0.70 to 0.91 and AUCs from 0.83 to 0.93. Correlations of sway with SARA were significant (maximum r = 0.65 and 0.73). Correlations with PROM ataxia were mild to moderate (maximum r = 0.56 and 0.34). Inertial sensor measures of extent of postural sway in conditions of natural stance EC and feet together stance EO were discriminative, reliable, and valid for monitoring SCA. © 2024 International Parkinson and Movement Disorder Society.

Exposures to Several Risk Factors can be Estimated from a Continuous Stream of Inertial Sensor Measurements during a Variety of Lifting-Lowering Tasks

Exposures to Several Risk Factors can be Estimated from a Continuous Stream of Inertial Sensor Measurements during a Variety of Lifting-Lowering Tasks

Experimental Validation of a Driver Monitoring System

This paper presents an analysis of the risk of neck injury in vehicle occupants as a consequence of an impact. A review of the formulation of indexes that are used in the assessment and investigation of neck injury risk is discussed with the aim of providing a new, more appropriate index using suitable sensorized equipment. An experimental analysis is proposed with a new driver monitoring device using low-cost sensors. The system consists of wearable units for the head, neck, and torso where inertial measurement sensors (IMU) are installed to record data concerning the occupant’s head, neck, and torso accelerations while the vehicle moves. Two laser infrared distance sensors are also installed on the vehicle’s steering wheel to record the position data of the head and neck, as well as an additional IMU for vehicle acceleration values. To validate both the device and the new index, experiments are designed in which different sensorized volunteers reproduce an emergency braking maneuver with an instrumented vehicle at speeds of 10, 20, and 30 km/h before the beginning of any braking action. The neck is particularly sensitive to sudden changes in acceleration, so a sudden braking maneuver is enough to constitute a potential risk of cervical spine injury. During the experiments, large accelerations and displacements were recorded as the test speed increased. The largest accelerations were obtained in the experimental test at a speed of 30 km/h with values of 19.17, 9.57, 9.28, and 5.09 m/s2 in the head, torso, neck, and vehicle, respectively. In the same experiment, the largest displacement of the head was 0.33 m and that of the neck was 0.27 m. Experimental results have verified that the designed device can be effectively used to characterize the biomechanical response of the neck in car impacts. The new index is also able to quantify a neck injury risk by taking into account the dynamics of a vehicle and the kinematics of the occupant’s head, neck, and torso. The numerical value of the new index is inversely proportional to the acceleration experienced by the vehicle occupant, so that small values indicate risky conditions.

Association between Self-reported Sleep Quality and Single-task Gait in Young Adults: A Study Using Machine Learning.

Objective The objective of the present study was to find biomechanical correlates of single-task gait and self-reported sleep quality in a healthy, young population by replicating a recently published study. Materials and Methods Young adults ( n = 123) were recruited and were asked to complete the Pittsburgh Sleep Quality Inventory to assess sleep quality. Gait variables ( n = 53) were recorded using a wearable inertial measurement sensor system on an indoor track. The data were split into training and test sets and then different machine learning models were applied. A post-hoc analysis of covariance (ANCOVA) was used to find statistically significant differences in gait variables between good and poor sleepers. Results AdaBoost models reported the highest correlation coefficient (0.77), with Support-Vector classifiers reporting the highest accuracy (62%). The most important features associated with poor sleep quality related to pelvic tilt and gait initiation. This indicates that overall poor sleepers have decreased pelvic tilt angle changes, specifically when initiating gait coming out of turns (first step pelvic tilt angle) and demonstrate difficulty maintaining gait speed. Discussion The results of the present study indicate that when using traditional gait variables, single-task gait has poor accuracy prediction for subjective sleep quality in young adults. Although the associations in the study are not as strong as those previously reported, they do provide insight into how gait varies in individuals who report poor sleep hygiene. Future studies should use larger samples to determine whether single task-gait may help predict objective measures of sleep quality especially in a repeated measures or longitudinal or intervention framework.

Smartphone MEMS Accelerometer and Gyroscope Measurement Errors: Laboratory Testing and Analysis of the Effects on Positioning Performance.

Embedding various sensors with powerful computing and storage capabilities in a small communication device, smartphones have become a prominent platform for navigation. With the increasing popularity of Apple CarPlay and Android Auto, smartphones are quickly replacing built-in automotive navigation solutions. On the other hand, smartphones are equipped with low-performance Micro Electro Mechanical Systems (MEMS) sensors to enhance their navigation performance in Global Navigation Satellite System (GNSS)-degraded or -denied environments. Compared with higher-grade inertial navigation systems (INS), MEMS-based INSs have a poor navigation performance due to large measurement errors. In this paper, we present laboratory test results on the stochastic and deterministic errors observed in MEMS inertial sensor measurements of five different smartphones from different manufacturers. Then, we describe and discuss the short-term effects of these errors on the pure inertial navigation performance and also on the navigation performance based on the tight coupling of INS with GNSS measurements using a smartphone.

Validity of an inertial measurement unit for the assessment of range and quality of movement during head and thoracic spine movements

BackgroundPatients with spinal pain often exhibit movement limitations and altered motor control, which can be challenging to measure accurately in clinical practice. Inertial measurement sensors present a promising new opportunity to develop valid, low-cost, and easy-to-use methods for assessing and monitoring spinal motion in a clinical setting. AimThis study aimed to investigate the agreement of an inertial sensor and a 3D camera system for assessing the range of motion (ROM) and quality of movement (QOM) in head and trunk single-plane movements. MethodsThirty-three healthy, pain-free volunteers were included. Each participant performed movements of the head (cervical flexion, extension, and lateral flexion) and trunk (trunk flexion, extension, rotation, and lateral flexion), which were simultaneously recorded by a 3D camera system and an inertial measurement unit (MOTI, Aalborg, Denmark). Agreement and consistency were analyzed for ROM and QOM by determining intraclass correlation coefficients (ICC), mean bias, and with Bland-Altman plots. ResultsThe agreement between systems was excellent for all movements (ICC between 0.91 and 1.00) for ROM and good to excellent for the QOM (ICC between 0.84 and 0.95). The mean bias for all movements (0.1–0.8°) was below the minimum acceptable difference between devices. The Bland-Altman plot indicated that MOTI systematically measured a slightly greater ROM and QOM than the 3D camera system for all neck and trunk movements. ConclusionThis study showed that MOTI is a feasible and potentially applicable option to assess ROM and QOM for head and trunk movements in experimental and clinical settings.

System-Identification-Based Activity Recognition Algorithms With Inertial Sensors.

This paper focuses on activity recognition using a single wearable inertial measurement sensor placed on the subject's chest. The ten activities that need to be identified include lying down, standing, sitting, bending and walking, among others. The activity recognition approach is based on using and identifying a transfer function associated with each activity. The appropriate input and output signals for each transfer function are first determined based on the norms of the sensor signals excited by that specific activity. Then the transfer function is identified using training data and a Wiener filter based on the auto-correlation and cross-correlation of the output and input signals. The activity occurring in real-time is recognized by computing and comparing the input-output errors associated with all the transfer functions. The performance of the developed system is evaluated using data from a group of Parkinson's disease subjects, including data obtained in a clinical setting and data obtained through remote home monitoring. On average, the developed system provides better than 90% accuracy in identifying each activity as it occurs. Activity recognition is particularly useful for PD patients in order to monitor their level of activity, characterize their postural instability and recognize high risk-activities in real-time that could lead to falls.

Effect of Fatigue on Movement Patterns During a Loaded Ruck March.

Loaded ruck marching is a common training and operational task for many members of the military. It is known to cause fatigue, affect soldier readiness, and can lead to traumatic and overuse injuries. Quantifying the gait changes that occur over the course of a loaded ruck march may provide a better understanding of injury mechanisms and potentially allow for development of individualized injury-prevention training programs. This study examined the change in soldiers' gait patterns over the course of a loaded ruck march in order to examine the correlation between fatigue and kinematic parameters. Fatigue is a subjective term that may encompass factors such as energy expenditure, muscle exhaustion, and cognitive engagement. Since it can be difficult to quantify, the current study makes the broad assumption that fatigue increases in some (potentially nonlinear) fashion during a loaded ruck march. Three platoons of soldiers participated in a field training exercise with inertial measurement sensors placed on their chests and ankles to record gait parameters throughout a 7-mile ruck march. The effects of fatigue on stride length, stride width, ankle yaw, and torso lean (anterior-posterior [AP] and side-to-side [SS]) were compared using one-way repeated measure analyses of variance. In comparing the first and last quarters of the ruck march, stride length decreased, stride width increased, stride width variability increased, AP torso lean variability increased, and SS torso lean variability increased. Although they do not describe a direct relationship to injury, these results can inform enhanced approaches to quantify and predict soldier fatigue and more reliably prevent future injury.

Robust state and protection-level estimation within tightly coupled GNSS/INS navigation system

In autonomous applications for mobility and transport, a high-rate and highly accurate vehicle-state estimation is achieved by fusing measurements of global navigation satellite systems (GNSS) and inertial sensors. The state estimation and its protection-level generation often suffer from satellite-signal disturbances in urban environments and subsequent poor parametrization of the satellite observables. Thus, we propose an innovative scheme involving an extended H∞ filter (EHF) for robust state estimation and zonotope for the protection-level generation. This scheme is shown as part of a tightly coupled navigation system based on an inertial navigation system and aided by the GPS/Galileo dual-constellation satellite navigation system. Specifically, GNSS pseudorange and deltarange observables are utilized. The experimental results of post-processing a real-world dataset show significant advantages of EHF against a conventional extended Kalman filter regarding the navigation accuracy and robustness under various GNSS measurement parametrizations and environmental circumstances. The zonotope-based protection-level calculation is proven valid, computationally affordable, and feasible for real-time implementations.

수평 위치정보 추정을 위한 관성/천측 항법시스템 설계 및 약결합/강결합 방식의 성능 비교

This paper describes a navigation system design for horizontal position estimation using inertial measurement sensors and celestial navigation. In space, stars are widely spread objects in the celestial sphere and have been used mainly to obtain attitude information through star observation. However, it is also possible to obtain information about the horizontal position with the altitude of the star. It is called celestial navigation which is the same principle that former navigators used to locate themselves while sailing on the sea. In particular, in deep space where GPS is not available, it is important to obtain information on the location by making use of stars that are relatively easy to observe. Therefore, we introduce a navigation system that can estimate horizontal position and design two types of systems, loosely coupled and tightly coupled depending on how the measurements are utilized. It is intended to help in the future design of navigation system using celestial navigation by simulation studies that not only verify whether the system correctly estimates horizontal position but also comparing the performance of loosely and tightly coupled methods.

Machine learning aided jump height estimate democratization through smartphone measures.

The peak height reached in a countermovement jump is a well established performance parameter. Its estimate is often entrusted to force platforms or body-worn inertial sensors. To date, smartphones may possibly be used as an alternative for estimating jump height, since they natively embed inertial sensors. For this purpose, 43 participants performed 4 countermovement jumps (172 in total) on two force platforms (gold standard). While jumping, participants held a smartphone in their hands, whose inertial sensor measures were recorded. After peak height was computed for both instrumentations, twenty-nine features were extracted, related to jump biomechanics and to signal time-frequency characteristics, as potential descriptors of soft tissues or involuntary arm swing artifacts. A training set (129 jumps - 75%) was created by randomly selecting elements from the initial dataset, the remaining ones being assigned to the test set (43 jumps - 25%). On the training set only, a Lasso regularization was applied to reduce the number of features, avoiding possible multicollinearity. A multi-layer perceptron with one hidden layer was trained for estimating the jump height from the reduced feature set. Hyperparameters optimization was performed on the multi-layer perceptron using a grid search approach with 5-fold cross validation. The best model was chosen according to the minimum negative mean absolute error. The multi-layer perceptron greatly improved the accuracy (4 cm) and precision (4 cm) of the estimates on the test set with respect to the raw smartphone measures estimates (18 and 16 cm, respectively). Permutation feature importance was performed on the trained model in order to establish the influence that each feature had on the outcome. The peak acceleration and the braking phase duration resulted the most influential features in the final model. Despite not being accurate enough, the height computed through raw smartphone measures was still among the most influential features. The study, implementing a smartphone-based method for jump height estimates, paves the way to method release to a broader audience, pursuing a democratization attempt.