[국내논문]Development of a Wearable Inertial Sensor-based Gait Analysis Device Using Machine Learning Algorithms -Validity of the Temporal Gait Parameter in Healthy Young Adults-원문보기
Purpose: The study aims were to develop a wearable inertial sensor-based gait analysis device that uses machine learning algorithms, and to validate this novel device using temporal gait parameters. Methods: Thirty-four healthy young participants (22 male, 12 female, aged 25.76 years) with no muscul...
Purpose: The study aims were to develop a wearable inertial sensor-based gait analysis device that uses machine learning algorithms, and to validate this novel device using temporal gait parameters. Methods: Thirty-four healthy young participants (22 male, 12 female, aged 25.76 years) with no musculoskeletal disorders were asked to walk at three different speeds. As they walked, data were simultaneously collected by a motion capture system and inertial measurement units (Reseed®). The data were sent to a machine learning algorithm adapted to the wearable inertial sensor-based gait analysis device. The validity of the newly developed instrument was assessed by comparing it to data from the motion capture system. Results: At normal speeds, intra-class correlation coefficients (ICC) for the temporal gait parameters were excellent (ICC [2, 1], 0.99~0.99), and coefficient of variation (CV) error values were insignificant for all gait parameters (0.31~1.08%). At slow speeds, ICCs for the temporal gait parameters were excellent (ICC [2, 1], 0.98~0.99), and CV error values were very small for all gait parameters (0.33~1.24%). At the fastest speeds, ICCs for temporal gait parameters were excellent (ICC [2, 1], 0.86~0.99) but less impressive than for the other speeds. CV error values were small for all gait parameters (0.17~5.58%). Conclusion: These results confirm that both the wearable inertial sensor-based gait analysis device and the machine learning algorithms have strong concurrent validity for temporal variables. On that basis, this novel wearable device is likely to prove useful for establishing temporal gait parameters while assessing gait.
Purpose: The study aims were to develop a wearable inertial sensor-based gait analysis device that uses machine learning algorithms, and to validate this novel device using temporal gait parameters. Methods: Thirty-four healthy young participants (22 male, 12 female, aged 25.76 years) with no musculoskeletal disorders were asked to walk at three different speeds. As they walked, data were simultaneously collected by a motion capture system and inertial measurement units (Reseed®). The data were sent to a machine learning algorithm adapted to the wearable inertial sensor-based gait analysis device. The validity of the newly developed instrument was assessed by comparing it to data from the motion capture system. Results: At normal speeds, intra-class correlation coefficients (ICC) for the temporal gait parameters were excellent (ICC [2, 1], 0.99~0.99), and coefficient of variation (CV) error values were insignificant for all gait parameters (0.31~1.08%). At slow speeds, ICCs for the temporal gait parameters were excellent (ICC [2, 1], 0.98~0.99), and CV error values were very small for all gait parameters (0.33~1.24%). At the fastest speeds, ICCs for temporal gait parameters were excellent (ICC [2, 1], 0.86~0.99) but less impressive than for the other speeds. CV error values were small for all gait parameters (0.17~5.58%). Conclusion: These results confirm that both the wearable inertial sensor-based gait analysis device and the machine learning algorithms have strong concurrent validity for temporal variables. On that basis, this novel wearable device is likely to prove useful for establishing temporal gait parameters while assessing gait.
* AI 자동 식별 결과로 적합하지 않은 문장이 있을 수 있으니, 이용에 유의하시기 바랍니다.
문제 정의
In conclusion, this study revealed the validity of the wearable inertial sensor on temporal parameters for gait analysis. It is thought that gait analysis will be more convenient under various conditions through a cost effective wearable inertial sensor.
제안 방법
Therefore, in this study, the data collected using the wearable IMU is analyzed by machine learning and the result is compared with the gait analysis result using the 3D gait analysis system to increase the validity of the temporal gait parameter through the wearable IMU.
For model tuning, the coarse-to-fine search method was used to find model parameters quickly over a wide range and reduce the range again to find the optimal model parameters. To train the model, we divided it into a training dataset and a testing dataset using 5-fold cross-validation. The model was trained using the training dataset and the model was validated with the testing dataset.
In this study, we developed a wearable inertial sensor-based gait analysis device using a machine learning algorithm and confirmed the validity of the temporal gait parameters’ evaluation.
The limitation of this study was that the experiment was conducted only for normal adults without spatial gait parameters. It is necessary to extend the validity of these findings by conducting research on participants with various diseases and from varying age groups.
대상 데이터
In a comparative study on the spatial and temporal gait analysis between the standard stationary treadmill and the inertial sensor attached to the body, which is commonly used in gait analysis, the speed and stride of slow gait in temporal gait characteristics It showed a slight difference except length. The study was conducted in healthy elderly people (Donath et al., 2016).
Thirty-four healthy young participants (25.76 ± 4.09 years old; 12 women, 22 men; height 170.14 ± 10.07cm;weight 66.47 ± 13.61kg) from S University in Seoul were recruited for this study.
For the infrared cameras to track motion reflective markers, clusters were attached to the thigh and shin to measure movement. To collect image data, four cameras were installed at the front, four cameras at the back and four cameras each on the left and right sides. All subjects wore leggings and a total of 52 markers were attached to the joints and segmental surfaces of the entire body and a resting calibration was performed.
To collect image data, four cameras were installed at the front, four cameras at the back and four cameras each on the left and right sides. All subjects wore leggings and a total of 52 markers were attached to the joints and segmental surfaces of the entire body and a resting calibration was performed. After removing the static marker, sufficient practice was conducted to induce a natural gait motion.
In addition, features with low importance were classified using a feature importance scale calculated by using the Light GBM algorithm. A total of 127 features were selected through this feature selection process.
To train the model, we divided it into a training dataset and a testing dataset using 5-fold cross-validation. The model was trained using the training dataset and the model was validated with the testing dataset. Jupyter notebook software (v.
데이터처리
The level of agreement between motion capture using an infrared camera and a wearable inertial sensor-based gait analysis device was analyzed using an in-class correlation coefficient (ICC [2, 1]) [21]. The coefficient of variation (CV) described by Bland and Altman, and 95% limits of agreement (LOA) were calculated to compare absolutely the parameters obtained in both sessions [23].
이론/모형
Because highly correlated features were duplicated, the performance of the model could have been affected thus requiring their removal. In addition, features with low importance were classified using a feature importance scale calculated by using the Light GBM algorithm. A total of 127 features were selected through this feature selection process.
For the machine learning stage of this study, a regression learning model was used. Gradient Boosting, one of the tree-based machine learning algorithms, was used to create a predictive model for regression analysis.
For the machine learning stage of this study, a regression learning model was used. Gradient Boosting, one of the tree-based machine learning algorithms, was used to create a predictive model for regression analysis.Before training the model, the training dataset was divided into training / validation datasets and the model was tuned.
The machine learning technique used in this study for the regression machine learning model, The Gradient Boosting Algorithm, is one of the tree-based machine learning algorithms, used to create a predictive model for regression analysis.
성능/효과
Many errors have been caused in using the data obtained from the wearable inertial sensor which attached to one side of the pelvis for evaluating gait parameters. However, using the machine learning technique produced a lower error rate and the utility of this equipment for gait analysis was confirmed.
참고문헌 (26)
Auvinet B, Berrut G, Touzard C, et al. Reference data for normal subjects obtained with an accelerometric device. Gait and Posture. 2002;16(2):124-134.
Bertoli M, Cereatti A, Trojaniello D, et al. Estimation of spatio-temporal parameters of gait from magnetoinertial measurement units: multicenter validation among Parkinson, mildly cognitively impaired and healthy older adults. Biomedical Engineering Online. 2018;17(1):58.
Bertuletti S, Cereatti A, Comotti D, et al. Static and dynamic accuracy of an innovative miniaturized wearable platform for short range distance measurements for human movement applications. Sensors (Basel). 2017;17(7):1492.
Caldas R, Mundt M, Potthast W, et al. A systematic review of gait analysis methods based on inertial sensors and adaptive algorithms. Gait Posture. 2017;57:204-210.
Chen S, Lach J, Lo B, et al. Toward pervasive gait analysis with wearable sensors: a systematic review. IEEE Journal of Biomedical and Health Informatics. 2016;20(6):1521-1537.
Donath L, Faude O, Lichtenstein E, et al. Validity and reliability of a portable gait analysis system for measuring spatiotemporal gait characteristics: comparison to an instrumented treadmill. Journal of Neuroengineering and Rehabilitation. 2016;13:6.
Donath L, Faude O, Lichtenstein E, et al. Mobile inertial sensor based gait analysis: validity and reliability of spatiotemporal gait characteristics in healthy seniors. Gait Posture. 2016;49:371-374.
Gonzalez-Recio O, Jimenez-Montero JAAlenda R. The gradient boosting algorithm and random boosting for genome-assisted evaluation in large data sets. Journal of Dairy Science. 2013;96(1):614-624.
Howcroft J, Kofman J, Lemaire ED, et al. Analysis of dual-task elderly gait in fallers and non-fallers using wearable sensors. Journal of Biomechanics. 2016;49(7):992-1001.
Kluge F, Gassner H, Hannink J, et al. Towards mobile gait analysis: concurrent validity and test-retest reliability of an inertial measurement system for the assessment of spatio-temporal gait parameters. Sensors (Basel). 2017;17(7):1622.
Kose A, Cereatti ADella Croce U. Bilateral step length estimation using a single inertial measurement unit attached to the pelvis. Journal of Neuroengineering and Rehabilitation. 2012;9:9.
Mancini MHorak FB. Potential o f APDM mobility lab for the monitoring of the progression of Parkinson's disease. Expert Review of Medical Devices. 2016;13(5):455-462.
Park JH, Mancini M, Carlson-Kuhta P, et al. Quantifying effects of age on balance and gait with inertial sensors in community-dwelling healthy adults. Experimental Gerontology. 2016;85:48-58.
Potluri S, Chandran AB, Diedrich C, et al. Machine learning based human gait segmentation with wearable sensor platform. Conference Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society. 2019;2019:588-594.
Rhudy MBMahoney JM. A comprehensive comparison of simple step counting techniques using wrist- and ankle-mounted accelerometer and gyroscope signals. Journal of Medical Engineering & Technology. 2018;42(3):236-243.
Senden R, Savelberg HH, Grimm B, et al. Accelerometry-based gait analysis, an additional objective approach to screen subjects at risk for falling. Gait and Posture. 2012;36(2):296-300.
Shull PB, Jirattigalachote W, Hunt MA, et al. Quantified self and human movement: a review on the clinical impact of wearable sensing and feedback for gait analysis and intervention. Gait and Posture. 2014;40(1):11-19.
Sposaro FTyson G. iFall: an android application for fall monitoring and response. Conference of the IEEE Engineering in Medicine and Biology Society. 2009;2009:6119-6122.
Takeda R, Lisco G, Fujisawa T, et al. Drift removal for improving the accuracy of gait parameters using wearable sensor systems. Sensors (Basel). 2014;14(12):23230-23247.
Teufl W, Lorenz M, Miezal M, et al. Towards inertial sensor based mobile gait analysis: event-detection and spatio-temporal parameters. Sensors (Basel). 2018;19(1):38.
Washabaugh EP, Kalyanaraman T, Adamczyk PG, et al. Validity and repeatability of inertial measurement units for measuring gait parameters. Gait Posture. 2017;55:87-93.
Yamada M, Aoyama T, Mori S, et al. Objective assessment of abnormal gait in patients with rheumatoid arthritis using a smartphone. Rheumatology International. 2012;32(12):3869-3874.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.