Wearable sensor validation of sports-related movements for the lower extremity and trunk

Dahl, Kimi D.; Dunford, Kristin M.; Wilson, Sandra Jo; Turnbull, Travis Lee; Tashman, Scott

doi:10.1016/j.medengphy.2020.08.001

Cited by 35 publications

(22 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This study used the root mean square error (RMSE) as the indicator to evaluate the prediction performance of each method as it is one of the most commonly used prediction performance indicators [88,89] and has been used as a standard statistical metric to measure model performance in meteorology, air quality, and climate researches [90][91][92]. Many sports outcome-prediction research used only RMSE as the prediction performance indicator [93][94][95]. Therefore, this research use RMSE as prediction performance indicator.…”

Section: Proposed Sports Outcome Prediction Processmentioning

confidence: 99%

Improving Sports Outcome Prediction Process Using Integrating Adaptive Weighted Features and Machine Learning Techniques

Lee

Wang

et al. 2021

Processes

View full text Add to dashboard Cite

Developing an effective sports performance analysis process is an attractive issue in sports team management. This study proposed an improved sports outcome prediction process by integrating adaptive weighted features and machine learning algorithms for basketball game score prediction. The feature engineering method is used to construct designed features based on game-lag information and adaptive weighting of variables in the proposed prediction process. These designed features are then applied to the five machine learning methods, including classification and regression trees (CART), random forest (RF), stochastic gradient boosting (SGB), eXtreme gradient boosting (XGBoost), and extreme learning machine (ELM) for constructing effective prediction models. The empirical results from National Basketball Association (NBA) data revealed that the proposed sports outcome prediction process could generate a promising prediction result compared to the competing models without adaptive weighting features. Our results also showed that the machine learning models with four game-lags information and adaptive weighting of power could generate better prediction performance.

show abstract

Section: Proposed Sports Outcome Prediction Processmentioning

confidence: 99%

Improving Sports Outcome Prediction Process Using Integrating Adaptive Weighted Features and Machine Learning Techniques

Lee

Wang

et al. 2021

Processes

View full text Add to dashboard Cite

show abstract

“…Each IMU consists of a tri-axial linear accelerometer, tri-axial rate gyroscopes that measure angular velocity, and a triaxial magnetometer to measure the orientation of the gyro [8]. The IMU is usually attached directly to the skin or via a Velcro or elastic strap or garment around the body segment [9,10]. IMUs have been incorporated into many wearable sensor systems in order to track joint kinematics during gait analysis, sports, and rehabilitation activities [9,11,12] because they have the advantage of being small, light, and easily attachable to different body segments.…”

Section: Introductionmentioning

confidence: 99%

“…When the IMU or marker is fastened directly to the skin, this vibration leads to a soft tissue motion artifact that degrades the kinematic recordings of what are meant to represent motions of the underlying bone [15,16]. Several authors have employed active or passive skin markers tracked by motion capture systems to evaluate the accuracy of measuring joint angles or joint angular velocities [10,13,17]. There have also been studies with magnetic resonance imaging (MRI) [18] and X-ray fluoroscopy measurements [15] of the kinematics of the underlying bone to remove the effect from soft tissue, but stereoradiography requires a static setup of the X-ray sources in the laboratory, is invasive and unethical for use on children or adolescents, and is completely impractical for use in the field.…”

Section: Introductionmentioning

confidence: 99%

A Comparison of Inertial Measurement Unit and Motion Capture Measurements of Tibiofemoral Kinematics during Simulated Pivot Landings

Baek

Ajdaroski

Shahshahani

et al. 2022

Sensors

View full text Add to dashboard Cite

Injuries are often associated with rapid body segment movements. We compared Certus motion capture and APDM inertial measurement unit (IMU) measurements of tibiofemoral angle and angular velocity changes during simulated pivot landings (i.e., ~70 ms peak) of nine cadaver knees dissected free of skin, subcutaneous fat, and muscle. Data from a total of 852 trials were compared using the Bland–Altman limits of agreement (LoAs): the Certus system was considered the gold standard measure for the angle change measurements, whereas the IMU was considered the gold standard for angular velocity changes. The results show that, although the mean peak IMU knee joint angle changes were slightly underestimated (2.1° for flexion, 0.2° for internal rotation, and 3.0° for valgus), the LoAs were large, ranging from 35.9% to 49.8%. In the case of the angular velocity changes, Certus had acceptable accuracy in the sagittal plane, with LoAs of ±54.9°/s and ±32.5°/s for the tibia and femur. For these rapid motions, we conclude that, even in the absence of soft tissues, the IMUs could not reliably measure these peak 3D knee angle changes; Certus measurements of peak tibiofemoral angular velocity changes depended on both the magnitude of the velocity and the plane of measurement.

show abstract

“…A gait detection device is proposed for lower-extremity exoskeleton robots, which is integrated with a smart sensor in the shoes and has a compact structure and strong practicability (Zeng et al, 2021). With the development of camera technology, machine vision technology is also introduced into the field of human limb rehabilitation field (Gherman et al, 2019;Dahl et al, 2020;Mavor et al, 2020). However, most of the human limb function evaluation systems based on machine vision require a combination of cameras.…”

Section: Introductionmentioning

confidence: 99%

Measurement Method of Human Lower Limb Joint Range of Motion Through Human-Machine Interaction Based on Machine Vision

et al. 2021

View full text Add to dashboard Cite

To provide stroke patients with good rehabilitation training, the rehabilitation robot should ensure that each joint of the limb of the patient does not exceed its joint range of motion. Based on the machine vision combined with an RGB-Depth (RGB-D) camera, a convenient and quick human-machine interaction method to measure the lower limb joint range of motion of the stroke patient is proposed. By analyzing the principle of the RGB-D camera, the transformation relationship between the camera coordinate system and the pixel coordinate system in the image is established. Through the markers on the human body and chair on the rehabilitation robot, an RGB-D camera is used to obtain their image data with relative position. The threshold segmentation method is used to process the image. Through the analysis of the image data with the least square method and the vector product method, the range of motion of the hip joint, knee joint in the sagittal plane, and hip joint in the coronal plane could be obtained. Finally, to verify the effectiveness of the proposed method for measuring the lower limb joint range of motion of human, the mechanical leg joint range of motion from a lower limb rehabilitation robot, which will be measured by the angular transducers and the RGB-D camera, was used as the control group and experiment group for comparison. The angle difference in the sagittal plane measured by the proposed detection method and angle sensor is relatively conservative, and the maximum measurement error is not more than 2.2 degrees. The angle difference in the coronal plane between the angle at the peak obtained by the designed detection system and the angle sensor is not more than 2.65 degrees. This paper provides an important and valuable reference for the future rehabilitation robot to set each joint range of motion limited in the safe workspace of the patient.

show abstract

Wearable sensor validation of sports-related movements for the lower extremity and trunk

Cited by 35 publications

References 19 publications

Improving Sports Outcome Prediction Process Using Integrating Adaptive Weighted Features and Machine Learning Techniques

Improving Sports Outcome Prediction Process Using Integrating Adaptive Weighted Features and Machine Learning Techniques

A Comparison of Inertial Measurement Unit and Motion Capture Measurements of Tibiofemoral Kinematics during Simulated Pivot Landings

Measurement Method of Human Lower Limb Joint Range of Motion Through Human-Machine Interaction Based on Machine Vision

Contact Info

Product

Resources

About