Sensors

Research

Jump to: Review

12 pages, 856 KiB

Open AccessArticle

Establishing Accelerometer Cut-Points to Classify Walking Speed in People Post Stroke

by David Moulaee Conradsson and Lucian John-Ross Bezuidenhout

Sensors 2022, 22(11), 4080; https://doi.org/10.3390/s22114080 - 27 May 2022

Cited by 1 | Viewed by 1746

Abstract

While accelerometers could be used to monitor important domains of walking in daily living (e.g., walking speed), the interpretation of accelerometer data often relies on validation studies performed with healthy participants. The aim of this study was to develop cut-points for waist- and [...] Read more.

While accelerometers could be used to monitor important domains of walking in daily living (e.g., walking speed), the interpretation of accelerometer data often relies on validation studies performed with healthy participants. The aim of this study was to develop cut-points for waist- and ankle-worn accelerometers to differentiate non-ambulation from walking and different walking speeds in people post stroke. Forty-two post-stroke persons wore waist and ankle accelerometers (ActiGraph GT3x+, AG) while performing three non-ambulation activities (i.e., sitting, setting the table and washing dishes) and while walking in self-selected and brisk speeds. Receiver operating characteristic (ROC) curve analysis was used to define AG cut-points for non-ambulation and different walking speeds (0.41–0.8 m/s, 0.81–1.2 m/s and >1.2 m/s) by considering sensor placement, axis, filter setting and epoch length. Optimal data input and sensor placements for measuring walking were a vector magnitude at 15 s epochs for waist- and ankle-worn AG accelerometers, respectively. Across all speed categories, cut-point classification accuracy was good-to-excellent for the ankle-worn AG accelerometer and fair-to-excellent for the waist-worn AG accelerometer, except for between 0.81 and 1.2 m/s. These cut-points can be used for investigating the link between walking and health outcomes in people post stroke. Full article

(This article belongs to the Special Issue Wearable and Integrated Sensors for Sport and Rehabilitation Applications)

► Show Figures

Figure 1

11 pages, 1656 KiB

Open AccessArticle

Stride Lengths during Maximal Linear Sprint Acceleration Obtained with Foot-Mounted Inertial Measurement Units

by Cornelis J. de Ruiter, Erik Wilmes, Pepijn S. van Ardenne, Niels Houtkamp, Reinder A. Prince, Maarten Wooldrik and Jaap H. van Dieën

Sensors 2022, 22(1), 376; https://doi.org/10.3390/s22010376 - 4 Jan 2022

Cited by 3 | Viewed by 2305

Abstract

Inertial measurement units (IMUs) fixed to the lower limbs have been reported to provide accurate estimates of stride lengths (SLs) during walking. Due to technical challenges, validation of such estimates in running is generally limited to speeds (well) below 5 m·s⁻¹. [...] Read more.

Inertial measurement units (IMUs) fixed to the lower limbs have been reported to provide accurate estimates of stride lengths (SLs) during walking. Due to technical challenges, validation of such estimates in running is generally limited to speeds (well) below 5 m·s⁻¹. However, athletes sprinting at (sub)maximal effort already surpass 5 m·s⁻¹ after a few strides. The present study aimed to develop and validate IMU-derived SLs during maximal linear overground sprints. Recreational athletes (n = 21) completed two sets of three 35 m sprints executed at 60, 80, and 100% of subjective effort, with an IMU on the instep of each shoe. Reference SLs from start to ~30 m were obtained with a series of video cameras. SLs from IMUs were obtained by double integration of horizontal acceleration with a zero-velocity update, corrected for acceleration artefacts at touch-down of the feet. Peak sprint speeds (mean ± SD) reached at the three levels of effort were 7.02 ± 0.80, 7.65 ± 0.77, and 8.42 ± 0.85 m·s⁻¹, respectively. Biases (±Limits of Agreement) of SLs obtained from all participants during sprints at 60, 80, and 100% effort were 0.01% (±6.33%), −0.75% (±6.39%), and −2.51% (±8.54%), respectively. In conclusion, in recreational athletes wearing IMUs tightly fixed to their shoes, stride length can be estimated with reasonable accuracy during maximal linear sprint acceleration. Full article

(This article belongs to the Special Issue Wearable and Integrated Sensors for Sport and Rehabilitation Applications)

► Show Figures

Figure 1

20 pages, 3798 KiB

Open AccessArticle

Assessing Visual Exploratory Activity of Athletes in Virtual Reality Using Head Motion Characteristics

by Markus Wirth, Sebastian Kohl, Stefan Gradl, Rosanna Farlock, Daniel Roth and Bjoern M. Eskofier

Sensors 2021, 21(11), 3728; https://doi.org/10.3390/s21113728 - 27 May 2021

Cited by 7 | Viewed by 3559

Abstract

Maximizing performance success in sports is about continuous learning and adaptation processes. Aside from physiological, technical and emotional performance factors, previous research focused on perceptual skills, revealing their importance for decision-making. This includes deriving relevant environmental information as a result of eye, head [...] Read more.

Maximizing performance success in sports is about continuous learning and adaptation processes. Aside from physiological, technical and emotional performance factors, previous research focused on perceptual skills, revealing their importance for decision-making. This includes deriving relevant environmental information as a result of eye, head and body movement interaction. However, to evaluate visual exploratory activity (VEA), generally utilized laboratory settings have restrictions that disregard the representativeness of assessment environments and/or decouple coherent cognitive and motor tasks. In vivo studies, however, are costly and hard to reproduce. Furthermore, the application of elaborate methods like eye tracking are cumbersome to implement and necessitate expert knowledge to interpret results correctly. In this paper, we introduce a virtual reality-based reproducible assessment method allowing the evaluation of VEA. To give insights into perceptual-cognitive processes, an easily interpretable head movement-based metric, quantifying VEA of athletes, is investigated. Our results align with comparable in vivo experiments and consequently extend them by showing the validity of the implemented approach as well as the use of virtual reality to determine characteristics among different skill levels. The findings imply that the developed method could provide accurate assessments while improving the control, validity and interpretability, which in turn informs future research and developments. Full article

(This article belongs to the Special Issue Wearable and Integrated Sensors for Sport and Rehabilitation Applications)

► Show Figures

Figure 1

23 pages, 1506 KiB

Open AccessEditor’s ChoiceArticle

From the Laboratory to the Field: IMU-Based Shot and Pass Detection in Football Training and Game Scenarios Using Deep Learning

by Maike Stoeve, Dominik Schuldhaus, Axel Gamp, Constantin Zwick and Bjoern M. Eskofier

Sensors 2021, 21(9), 3071; https://doi.org/10.3390/s21093071 - 28 Apr 2021

Cited by 22 | Viewed by 4336

Abstract

The applicability of sensor-based human activity recognition in sports has been repeatedly shown for laboratory settings. However, the transferability to real-world scenarios cannot be granted due to limitations on data and evaluation methods. On the example of football shot and pass detection against [...] Read more.

The applicability of sensor-based human activity recognition in sports has been repeatedly shown for laboratory settings. However, the transferability to real-world scenarios cannot be granted due to limitations on data and evaluation methods. On the example of football shot and pass detection against a null class we explore the influence of those factors for real-world event classification in field sports. For this purpose we compare the performance of an established Support Vector Machine (SVM) for laboratory settings from literature to the performance in three evaluation scenarios gradually evolving from laboratory settings to real-world scenarios. In addition, three different types of neural networks, namely a convolutional neural net (CNN), a long short term memory net (LSTM) and a convolutional LSTM (convLSTM) are compared. Results indicate that the SVM is not able to reliably solve the investigated three-class problem. In contrast, all deep learning models reach high classification scores showing the general feasibility of event detection in real-world sports scenarios using deep learning. The maximum performance with a weighted f1-score of 0.93 was reported by the CNN. The study provides valuable insights for sports assessment under practically relevant conditions. In particular, it shows that (1) the discriminative power of established features needs to be reevaluated when real-world conditions are assessed, (2) the selection of an appropriate dataset and evaluation method are both required to evaluate real-world applicability and (3) deep learning-based methods yield promising results for real-world HAR in sports despite high variations in the execution of activities. Full article

(This article belongs to the Special Issue Wearable and Integrated Sensors for Sport and Rehabilitation Applications)

► Show Figures

Figure 1

15 pages, 3267 KiB

Open AccessArticle

Bicycling Phase Recognition for Lower Limb Amputees Using Support Vector Machine Optimized by Particle Swarm Optimization

by Xinxin Li, Zuojun Liu, Xinzhi Gao and Jie Zhang

Sensors 2020, 20(22), 6533; https://doi.org/10.3390/s20226533 - 15 Nov 2020

Cited by 3 | Viewed by 3221

Abstract

A novel method for recognizing the phases in bicycling of lower limb amputees using support vector machine (SVM) optimized by particle swarm optimization (PSO) is proposed in this paper. The method is essential for enhanced prosthetic knee joint control for lower limb amputees [...] Read more.

A novel method for recognizing the phases in bicycling of lower limb amputees using support vector machine (SVM) optimized by particle swarm optimization (PSO) is proposed in this paper. The method is essential for enhanced prosthetic knee joint control for lower limb amputees in carrying out bicycling activity. Some wireless wearable accelerometers and a knee joint angle sensor are installed in the prosthesis to obtain data on the knee joint and ankle joint horizontal, vertical acceleration signal and knee joint angle. In order to overcome the problem of high noise content in the collected data, a soft-hard threshold filter was used to remove the noise caused by the vibration. The filtered information is then used to extract the multi-dimensional feature vector for the training of SVM for performing bicycling phase recognition. The SVM is optimized by PSO to enhance its classification accuracy. The recognition accuracy of the PSO-SVM classification model on testing data is 93%, which is much higher than those of BP, SVM and PSO-BP classification models. Full article

(This article belongs to the Special Issue Wearable and Integrated Sensors for Sport and Rehabilitation Applications)

► Show Figures

Figure 1

21 pages, 6460 KiB

Open AccessArticle

Validation of Spatiotemporal and Kinematic Measures in Functional Exercises Using a Minimal Modeling Inertial Sensor Methodology

by Benjamin R. Hindle, Justin W.L. Keogh and Anna V. Lorimer

Sensors 2020, 20(16), 4586; https://doi.org/10.3390/s20164586 - 15 Aug 2020

Cited by 9 | Viewed by 3161

Abstract

This study proposes a minimal modeling magnetic, angular rate and gravity (MARG) methodology for assessing spatiotemporal and kinematic measures of functional fitness exercises. Thirteen healthy persons performed repetitions of the squat, box squat, sandbag pickup, shuffle-walk, and bear crawl. Sagittal plane hip, knee, [...] Read more.

This study proposes a minimal modeling magnetic, angular rate and gravity (MARG) methodology for assessing spatiotemporal and kinematic measures of functional fitness exercises. Thirteen healthy persons performed repetitions of the squat, box squat, sandbag pickup, shuffle-walk, and bear crawl. Sagittal plane hip, knee, and ankle range of motion (ROM) and stride length, stride time, and stance time measures were compared for the MARG method and an optical motion capture (OMC) system. The root mean square error (RMSE), mean absolute percentage error (MAPE), and Bland–Altman plots and limits of agreement were used to assess agreement between methods. Hip and knee ROM showed good to excellent agreement with the OMC system during the squat, box squat, and sandbag pickup (RMSE: 4.4–9.8°), while ankle ROM agreement ranged from good to unacceptable (RMSE: 2.7–7.2°). Unacceptable hip and knee ROM agreement was observed for the shuffle-walk and bear crawl (RMSE: 3.3–8.6°). The stride length, stride time, and stance time showed good to excellent agreement between methods (MAPE: (3.2 ± 2.8)%–(8.2 ± 7.9)%). Although the proposed MARG-based method is a valid means of assessing spatiotemporal and kinematic measures during various exercises, further development is required to assess the joint kinematics of small ROM, high velocity movements. Full article

(This article belongs to the Special Issue Wearable and Integrated Sensors for Sport and Rehabilitation Applications)

► Show Figures

Figure 1

15 pages, 2074 KiB

Open AccessArticle

Golf Swing Segmentation from a Single IMU Using Machine Learning

by Myeongsub Kim and Sukyung Park

Sensors 2020, 20(16), 4466; https://doi.org/10.3390/s20164466 - 10 Aug 2020

Cited by 28 | Viewed by 7124

Abstract

Golf swing segmentation with inertial measurement units (IMUs) is an essential process for swing analysis using wearables. However, no attempt has been made to apply machine learning models to estimate and divide golf swing phases. In this study, we proposed and verified two [...] Read more.

Golf swing segmentation with inertial measurement units (IMUs) is an essential process for swing analysis using wearables. However, no attempt has been made to apply machine learning models to estimate and divide golf swing phases. In this study, we proposed and verified two methods using machine learning models to segment the full golf swing into five major phases, including before and after the swing, from every single IMU attached to a body part. Proposed bidirectional long short-term memory-based and convolutional neural network-based methods rely on characteristics that automatically learn time-series features, including sequential body motion during a golf swing. Nine professional and eleven skilled male golfers participated in the experiment to collect swing data for training and verifying the methods. We verified the proposed methods using leave-one-out cross-validation. The results revealed average segmentation errors of 5–92 ms from each IMU attached to the head, wrist, and waist, accurate compared to the heuristic method in this study. In addition, both proposed methods could segment all the swing phases using only the acceleration data, bringing advantage in terms of power consumption. This implies that swing-segmentation methods using machine learning could be applied to various motion-analysis environments by dividing motion phases with less restriction on IMU placement. Full article

(This article belongs to the Special Issue Wearable and Integrated Sensors for Sport and Rehabilitation Applications)

► Show Figures

Figure 1

Review

Jump to: Research

33 pages, 2878 KiB

Open AccessReview

Review of Real-Time Biomechanical Feedback Systems in Sport and Rehabilitation

by Matevž Hribernik, Anton Umek, Sašo Tomažič and Anton Kos

Sensors 2022, 22(8), 3006; https://doi.org/10.3390/s22083006 - 14 Apr 2022

Cited by 15 | Viewed by 4323

Abstract

Real-time biomechanical feedback (BMF) is a relatively new area of research. The potential of using advanced technology to improve motion skills in sport and accelerate physical rehabilitation has been demonstrated in a number of studies. This paper provides a literature review of BMF [...] Read more.

Real-time biomechanical feedback (BMF) is a relatively new area of research. The potential of using advanced technology to improve motion skills in sport and accelerate physical rehabilitation has been demonstrated in a number of studies. This paper provides a literature review of BMF systems in sports and rehabilitation. Our motivation was to examine the history of the field to capture its evolution over time, particularly how technologies are used and implemented in BMF systems, and to identify the most recent studies showing novel solutions and remarkable implementations. We searched for papers in three research databases: Scopus, Web of Science, and PubMed. The initial search yielded 1167 unique papers. After a rigorous and challenging exclusion process, 144 papers were eventually included in this report. We focused on papers describing applications and systems that implement a complete real-time feedback loop, which must include the use of sensors, real-time processing, and concurrent feedback. A number of research questions were raised, and the papers were studied and evaluated accordingly. We identified different types of physical activities, sensors, modalities, actuators, communications, settings and end users. A subset of the included papers, showing the most perspectives, was reviewed in depth to highlight and present their innovative research approaches and techniques. Real-time BMF has great potential in many areas. In recent years, sensors have been the main focus of these studies, but new types of processing devices, methods, and algorithms, actuators, and communication technologies and protocols will be explored in more depth in the future. This paper presents a broad insight into the field of BMF. Full article

(This article belongs to the Special Issue Wearable and Integrated Sensors for Sport and Rehabilitation Applications)

► Show Figures