Sensors

Research

15 pages, 2866 KiB

Open AccessArticle

Evaluating Sparse Inertial Measurement Unit Configurations for Inferring Treadmill Running Motion

by Mackenzie N. Pitts, Megan R. Ebers, Cristine E. Agresta and Katherine M. Steele

Sensors 2025, 25(7), 2105; https://doi.org/10.3390/s25072105 - 27 Mar 2025

Viewed by 254

Inertial measurement units (IMUs) are used to analyze running performance. While leveraging one sensor to estimate kinematic and kinetic variables is common, sparsity limits the number of digital biomarkers that can be evaluated. Shallow recurrent decoder networks (SHRED) can reconstruct a dense set [...] Read more.

Inertial measurement units (IMUs) are used to analyze running performance. While leveraging one sensor to estimate kinematic and kinetic variables is common, sparsity limits the number of digital biomarkers that can be evaluated. Shallow recurrent decoder networks (SHRED) can reconstruct a dense set of time-series signals from a single input sensor and have been successful in human mobility applications, highlighting the potential for this algorithm to monitor running. We trained and tested subject-specific SHRED models of nine subjects running on a treadmill to map from one input sensor to the remaining three IMUs. We varied the type of input to reflect experimental parameters that are important in running studies—sensor location, sensor type, sampling rate, and running speed—and compared the error of inferred signals from each input type. Sensor location and type did not impact SHRED inference accuracy, while decreasing the sampling rate affected the accuracy of ankle measurements. All ankle acceleration inferences from these models remained below the minimal detectable change threshold of 12.0 m/s². SHRED models trained and tested at multiple speeds did not accurately infer IMU measurements below this threshold. SHRED may broaden the scope of motion analysis by expanding datasets with fewer sensors. Full article

(This article belongs to the Special Issue Wearable Sensors and Machine Learning in the Biomechanics of Human Movement)

► Show Figures

Figure 1

20 pages, 8511 KiB

Open AccessArticle

Prediction of Vertical Ground Reaction Forces Under Different Running Speeds: Integration of Wearable IMU with CNN-xLSTM

by Tianxiao Chen, Datao Xu, Zhifeng Zhou, Huiyu Zhou, Shirui Shao and Yaodong Gu

Sensors 2025, 25(4), 1249; https://doi.org/10.3390/s25041249 - 18 Feb 2025

Viewed by 617

Abstract

Traditional methods for collecting ground reaction forces (GRFs) mainly use lab force plates. Previous research broke this pattern by predicting GRFs with deep learning and data from IMUs like joint acceleration. Joint angle, as a geometric, is easier to collect than acceleration outdoors [...] Read more.

Traditional methods for collecting ground reaction forces (GRFs) mainly use lab force plates. Previous research broke this pattern by predicting GRFs with deep learning and data from IMUs like joint acceleration. Joint angle, as a geometric, is easier to collect than acceleration outdoors with cameras. LSTM is one of the deep learning models that have shown good performance in biomechanical studies. xLSTM, as an optimized version of LSTM, has not been used in biomechanical studies and no research has predicted GRFs during running solely using lower limb joint angles. This study collected lower-limb joint angle and vertical ground reaction force data at five speeds from 12 healthy male runners with Xsens sensors. Datasets including three joints and three planes were set as the inputs of four deep learning models for vertical-GRF prediction. CNN-xLSTM consistently performed best in the four deep learning models when different datasets were input (R² = 0.909 ± 0.064, MAPE = 2.18 ± 0.09, rMSE = 0.061 ± 0.008), and the performance was at a relatively high level at the five speeds. The current findings may contribute to a new GRF measurement and provide a reference for future real-time motion detection and sport injury prediction. Full article

(This article belongs to the Special Issue Wearable Sensors and Machine Learning in the Biomechanics of Human Movement)

► Show Figures

Figure 1

18 pages, 2412 KiB

Open AccessArticle

Infants Display Anticipatory Gaze During a Motor Contingency Paradigm

by Marcelo R. Rosales, José Carlos Pulido, Carolee Winstein, Nina S. Bradley, Maja Matarić and Beth A. Smith

Sensors 2025, 25(3), 844; https://doi.org/10.3390/s25030844 - 30 Jan 2025

Viewed by 796

Abstract

Background: Examining visual behavior during a motor learning paradigm can enhance our understanding of how infants learn motor skills. The aim of this study was to determine if infants who learned a contingency visually anticipated the outcomes of their behavior. Methods: 15 infants [...] Read more.

Background: Examining visual behavior during a motor learning paradigm can enhance our understanding of how infants learn motor skills. The aim of this study was to determine if infants who learned a contingency visually anticipated the outcomes of their behavior. Methods: 15 infants (6–9 months of age) participated in a contingency learning paradigm. When an infant produced a right leg movement, a robot provided reinforcement by clapping. Three types of visual gaze events were identified: predictive, reactive, and not looking. An exploratory analysis examined the trends in visual-motor behavior that can be used to inform future questions and practices in contingency learning studies. Results: All classically defined learners visually anticipated robot activation at greater than random chance (W = 21; p = 0.028). Specifically, all but one learners displayed a distribution of gaze timing identified as predictive (skewness: 0.56–2.42) with the median timing preceding robot activation by 0.31 s (range: −0.40–0.18 s). Conclusions: Findings suggest that most learners displayed visual anticipation withing the first minutes of performing the paradigm. Further, the classical definition of learning a contingency paradigm in infants can be sharpened to further the design of contingency learning studies and advance the processes infants use to learn motor skills. Full article

(This article belongs to the Special Issue Wearable Sensors and Machine Learning in the Biomechanics of Human Movement)

► Show Figures

Figure 1

17 pages, 791 KiB

Open AccessArticle

Using Deep Learning Models to Predict Prosthetic Ankle Torque

by Christopher Prasanna, Jonathan Realmuto, Anthony Anderson, Eric Rombokas and Glenn Klute

Sensors 2023, 23(18), 7712; https://doi.org/10.3390/s23187712 - 6 Sep 2023

Cited by 4 | Viewed by 2548

Abstract

Inverse dynamics from motion capture is the most common technique for acquiring biomechanical kinetic data. However, this method is time-intensive, limited to a gait laboratory setting, and requires a large array of reflective markers to be attached to the body. A practical alternative [...] Read more.

Inverse dynamics from motion capture is the most common technique for acquiring biomechanical kinetic data. However, this method is time-intensive, limited to a gait laboratory setting, and requires a large array of reflective markers to be attached to the body. A practical alternative must be developed to provide biomechanical information to high-bandwidth prosthesis control systems to enable predictive controllers. In this study, we applied deep learning to build dynamical system models capable of accurately estimating and predicting prosthetic ankle torque from inverse dynamics using only six input signals. We performed a hyperparameter optimization protocol that automatically selected the model architectures and learning parameters that resulted in the most accurate predictions. We show that the trained deep neural networks predict ankle torques one sample into the future with an average RMSE of 0.04 ± 0.02 Nm/kg, corresponding to 2.9 ± 1.6% of the ankle torque’s dynamic range. Comparatively, a manually derived analytical regression model predicted ankle torques with a RMSE of 0.35 ± 0.53 Nm/kg, corresponding to 26.6 ± 40.9% of the ankle torque’s dynamic range. In addition, the deep neural networks predicted ankle torque values half a gait cycle into the future with an average decrease in performance of 1.7% of the ankle torque’s dynamic range when compared to the one-sample-ahead prediction. This application of deep learning provides an avenue towards the development of predictive control systems for powered limbs aimed at optimizing prosthetic ankle torque. Full article

(This article belongs to the Special Issue Wearable Sensors and Machine Learning in the Biomechanics of Human Movement)

► Show Figures

Figure 1

10 pages, 3081 KiB

Open AccessArticle

Evaluation of a Restoration Algorithm Applied to Clipped Tibial Acceleration Signals

by Zoe Y. S. Chan, Chloe Angel, Daniel Thomson, Reed Ferber, Sharon M. H. Tsang and Roy T. H. Cheung

Sensors 2023, 23(10), 4609; https://doi.org/10.3390/s23104609 - 10 May 2023

Viewed by 1857

Abstract

Wireless accelerometers with various operating ranges have been used to measure tibial acceleration. Accelerometers with a low operating range output distorted signals and have been found to result in inaccurate measurements of peaks. A restoration algorithm using spline interpolation has been proposed to [...] Read more.

Wireless accelerometers with various operating ranges have been used to measure tibial acceleration. Accelerometers with a low operating range output distorted signals and have been found to result in inaccurate measurements of peaks. A restoration algorithm using spline interpolation has been proposed to restore the distorted signal. This algorithm has been validated for axial peaks within the range of 15.0–15.9 g. However, the accuracy of peaks of higher magnitude and the resultant peaks have not been reported. The purpose of the present study is to evaluate the measurement agreement of the restored peaks using a low-range accelerometer (±16 g) against peaks sampled using a high-range accelerometer (±200 g). The measurement agreement of both the axial and resultant peaks were examined. In total, 24 runners were equipped with 2 tri-axial accelerometers at their tibia and completed an outdoor running assessment. The accelerometer with an operating range of ±200 g was used as reference. The results of this study showed an average difference of −1.40 ± 4.52 g and −1.23 ± 5.48 g for axial and resultant peaks. Based on our findings, the restoration algorithm could skew data and potentially lead to incorrect conclusions if used without caution. Full article

(This article belongs to the Special Issue Wearable Sensors and Machine Learning in the Biomechanics of Human Movement)

► Show Figures

Figure 1

18 pages, 3106 KiB

Open AccessArticle

How Accurately Can Wearable Sensors Assess Low Back Disorder Risks during Material Handling? Exploring the Fundamental Capabilities and Limitations of Different Sensor Signals

by Cameron A. Nurse, Laura Jade Elstub, Peter Volgyesi and Karl E. Zelik

Sensors 2023, 23(4), 2064; https://doi.org/10.3390/s23042064 - 12 Feb 2023

Cited by 7 | Viewed by 2696

Abstract

Low back disorders (LBDs) are a leading occupational health issue. Wearable sensors, such as inertial measurement units (IMUs) and/or pressure insoles, could automate and enhance the ergonomic assessment of LBD risks during material handling. However, much remains unknown about which sensor signals to [...] Read more.

Low back disorders (LBDs) are a leading occupational health issue. Wearable sensors, such as inertial measurement units (IMUs) and/or pressure insoles, could automate and enhance the ergonomic assessment of LBD risks during material handling. However, much remains unknown about which sensor signals to use and how accurately sensors can estimate injury risk. The objective of this study was to address two open questions: (1) How accurately can we estimate LBD risk when combining trunk motion and under-the-foot force data (simulating a trunk IMU and pressure insoles used together)? (2) How much greater is this risk assessment accuracy than using only trunk motion (simulating a trunk IMU alone)? We developed a data-driven simulation using randomized lifting tasks, machine learning algorithms, and a validated ergonomic assessment tool. We found that trunk motion-based estimates of LBD risk were not strongly correlated (r range: 0.20–0.56) with ground truth LBD risk, but adding under-the-foot force data yielded strongly correlated LBD risk estimates (r range: 0.93–0.98). These results raise questions about the adequacy of a single IMU for LBD risk assessment during material handling but suggest that combining an IMU on the trunk and pressure insoles with trained algorithms may be able to accurately assess risks. Full article

(This article belongs to the Special Issue Wearable Sensors and Machine Learning in the Biomechanics of Human Movement)

► Show Figures

Figure 1

14 pages, 11408 KiB

Open AccessArticle

A Quantitative Assessment Grading Study of Balance Performance Based on Lower Limb Dataset

by Fei Wang, Anqi Dong, Kaiyu Zhang, Dexing Qian and Yinsheng Tian

Sensors 2023, 23(1), 33; https://doi.org/10.3390/s23010033 - 20 Dec 2022

Cited by 3 | Viewed by 2037

Abstract

Balance ability is one of the important factors in measuring human physical fitness and a common index for evaluating sports performance. Its quality directly affects the coordination ability of human movements and plays an important role in human productive activities. In the field [...] Read more.

Balance ability is one of the important factors in measuring human physical fitness and a common index for evaluating sports performance. Its quality directly affects the coordination ability of human movements and plays an important role in human productive activities. In the field of sports, balance ability is an important indicator of athletes’ selection and training. How to objectively analyze balance performance becomes a problem for every non-professional sports enthusiast. Therefore, in this paper, we used a dataset of lower limb collected by inertial sensors to extract the feature parameters, then designed a RUS Boost classifier for unbalanced data whose basic classifier was SVM model to predict three classifications of balance degree, and, finally, evaluated the performance of the new classifier by comparing it with two basic classifiers (KNN, SVM). The result showed that the new classifier could be used to evaluate the balanced ability of lower limb, and performed higher than basic ones (RUS Boost: 72%; KNN: 60%; SVM: 44%). The results meant the established classification model could be used for and quantitative assessment of balance ability in initial screening and targeted training. Full article

(This article belongs to the Special Issue Wearable Sensors and Machine Learning in the Biomechanics of Human Movement)

► Show Figures

Figure 1

15 pages, 2674 KiB

Open AccessArticle

Ambulation Mode Classification of Individuals with Transfemoral Amputation through A-Mode Sonomyography and Convolutional Neural Networks

by Rosemarie Murray, Joel Mendez, Lukas Gabert, Nicholas P. Fey, Honghai Liu and Tommaso Lenzi

Sensors 2022, 22(23), 9350; https://doi.org/10.3390/s22239350 - 1 Dec 2022

Cited by 11 | Viewed by 2766

Abstract

Many people struggle with mobility impairments due to lower limb amputations. To participate in society, they need to be able to walk on a wide variety of terrains, such as stairs, ramps, and level ground. Current lower limb powered prostheses require different control [...] Read more.

Many people struggle with mobility impairments due to lower limb amputations. To participate in society, they need to be able to walk on a wide variety of terrains, such as stairs, ramps, and level ground. Current lower limb powered prostheses require different control strategies for varying ambulation modes, and use data from mechanical sensors within the prosthesis to determine which ambulation mode the user is in. However, it can be challenging to distinguish between ambulation modes. Efforts have been made to improve classification accuracy by adding electromyography information, but this requires a large number of sensors, has a low signal-to-noise ratio, and cannot distinguish between superficial and deep muscle activations. An alternative sensing modality, A-mode ultrasound, can detect and distinguish between changes in superficial and deep muscles. It has also shown promising results in upper limb gesture classification. Despite these advantages, A-mode ultrasound has yet to be employed for lower limb activity classification. Here we show that A- mode ultrasound can classify ambulation mode with comparable, and in some cases, superior accuracy to mechanical sensing. In this study, seven transfemoral amputee subjects walked on an ambulation circuit while wearing A-mode ultrasound transducers, IMU sensors, and their passive prosthesis. The circuit consisted of sitting, standing, level-ground walking, ramp ascent, ramp descent, stair ascent, and stair descent, and a spatial–temporal convolutional network was trained to continuously classify these seven activities. Offline continuous classification with A-mode ultrasound alone was able to achieve an accuracy of

91.8 \pm 3.4 %

, compared with

93.8 \pm 3.0 %

, when using kinematic data alone. Combined kinematic and ultrasound produced

95.8 \pm 2.3 %

accuracy. This suggests that A-mode ultrasound provides additional useful information about the user’s gait beyond what is provided by mechanical sensors, and that it may be able to improve ambulation mode classification. By incorporating these sensors into powered prostheses, users may enjoy higher reliability for their prostheses, and more seamless transitions between ambulation modes. Full article

(This article belongs to the Special Issue Wearable Sensors and Machine Learning in the Biomechanics of Human Movement)

► Show Figures

Figure 1

Journal Menu

Journal Browser

Wearable Sensors and Machine Learning in the Biomechanics of Human Movement

Share This Special Issue

Special Issue Editors

Special Issue Information

Keywords

Benefits of Publishing in a Special Issue

Published Papers (8 papers)

Research

Further Information

Guidelines

MDPI Initiatives

Follow MDPI