Please use this identifier to cite or link to this item:
標題: A novel approach for assessing walk speed using waist mounted accelerometers
作者: 陳與祥
Yu-Shiang Chen
關鍵字: 逐步迴歸分析
Stepwise Regression Analysis
Root Mean Square error
引用: [1] Meijer, G. A., Westerterp, K. R., Koper, H., and ten Hoor, F. (1989). Assessment of energy expenditure by recording heart rate and body acceleration. Medicine and science in sports and exercise, 21(3), 343-347. [2] Aminian, K., De Andres, E., Rezakhanlou, K., Fritsch, C., Schutz, Y., Depairon, M., ... & Robert, P. (1998). Motion analysis in clinical practice using ambulatory accelerometry. In Modelling and Motion Capture Techniques for Virtual Environments (pp. 1-11). Springer Berlin Heidelberg. [3] Song, Y., Shin, S., Kim, S., Lee, D., & Lee, K. H. (2007, August). Speed estimation from a tri-axial accelerometer using neural networks. In Engineering in Medicine and Biology Society, 2007. EMBS 2007. 29th Annual International Conference of the IEEE (pp. 3224-3227). IEEE. [4] Aminian, K., Robert, P., Jequier, E., & Schutz, Y. (1995). Estimation of speed and incline of walking using neural network. Instrumentation and Measurement, IEEE Transactions on, 44(3), 743-746. [5] Weinberg, H. (2002). Using the ADXL202 in pedometer and personal navigation applications. Analog Devices AN-602 application note. [6] Schutz, Y., Weinsier, S., Terrier, P., & Durrer, D. (2002). A new accelerometric method to assess the daily walking practice. International journal of obesity and related metabolic disorders: journal of the International Association for the Study of Obesity, 26(1), 111-118. [7] Freedson, P. S., & Miller, K. (2000). Objective monitoring of physical activity using motion sensors and heart rate. Research Quarterly for Exercise and Sport, 71(2 Suppl), S21-9. [8] Hendelman, D., Miller, K., Baggett, C., Debold, E., & Freedson, P. (2000). Validity of accelerometry for the assessment of moderate intensity physical activity in the field. Medicine and science in sports and exercise, 32(9 Suppl), S442-9. [9] Fehling, P. C., Smith, D. L., Warner, S. E., & Dalsky, G. P. (1999). Comparison of accelerometers with oxygen consumption in older adults during exercise. Medicine and science in sports and exercise, 31(1), 171-175. [10] Terrier, P., Aminian, K., & Schutz, Y. (2001). Can accelerometry accurately predict the energy cost of uphill/downhill walking?. Ergonomics, 44(1), 48-62. [11] Ayen, T. G., 'Estimation of energy expenditure with a simulated three-dimensional accelerometer,' Ambulatory Monitoring, vol. 1, pp.293-301, 1988. [12] Montoye, H. J., Washburn, R., Servais, S., Ertl, A., Webster, J. G., and Nagle, F. J., 'Estimation of energy expenditure by a portable accelerometer,' Med Sci Sports Exerc, vol. 15, pp. 403-7, 1983. [13] Yoshida, T., Sakane, N., Umekawa, T., & Kondo, M. (1994). Relationship between basal metabolic rate, thermogenic response to caffeine, and body weight loss following combined low calorie and exercise treatment in obese women. International journal of obesity and related metabolic disorders: journal of the International Association for the Study of Obesity, 18(5), 345-350. [14] 林其範、高材,'以三軸加速規和心率估測日常活動能量消耗',國立陽明大學,碩士論文,民國97年10月。 [15] He, Z., & Zhang, W. (2011). Estimation of walking speed using accelerometer and artificial neural networks. In Computer Science for Environmental Engineering and EcoInformatics (pp. 42-47). Springer Berlin Heidelberg. [16] Li, Q., Young, M., Naing, V., & Donelan, J. M. (2009, June). Walking speed and slope estimation using shank-mounted inertial measurement units. InRehabilitation Robotics, 2009. ICORR 2009. IEEE International Conference on(pp. 839-844). IEEE. [17] Sayeed, T., Sama, A., Catala, A., & Cabestany, J. (2013, September). Comparison and adaptation of step length and gait speed estimators from single belt worn accelerometer positioned on lateral side of the body. In Intelligent Signal Processing (WISP), 2013 IEEE 8th International Symposium on (pp. 14-20). IEEE. [18] Martin T. Hagan, Howard B. Dcmuth, Mark Beale, 'Neural Network Design'(2002) [19] Kavanagh, J. J., Morrison, S., James, D. A., & Barrett, R. (2006). Reliability of segmental accelerations measured using a new wireless gait analysis system.Journal of Biomechanics, 39(15), 2863-2872. [20] Meijer, G. A., Westerterp, K. R., Koper, H. A. N. S., & ten Hoor, F. O. P. P. E. (1989). Assessment of energy expenditure by recording heart rate and body acceleration. Medicine and science in sports and exercise, 21(3), 343-347. [21] Wong, T. C., Webster, J. G., Montoye, H. J., & Washburn, R. (1981). Portable accelerometer device for measuring human energy expenditure. Biomedical Engineering, IEEE Transactions on, (6), 467-471.
摘要: In this study, we developed a speed estimation model with human body accelerations measured on the waist by a tri-axial accelerometer. To estimate the speed we applied neural network. The sampling frequency is at 200 Hz. A total of 53 subjects (34 males and 19 females, age 22.2±1.32 years, height 170.4±7.7 cm) with the tri-axial accelerometer in their waist and walked 50 meters long hallway in free-living conditions. Each subject's actual average speed were count step by step. We extract eight parameters, namely step number, subject's height, root mean square (RMS) and difference between the maximum and minimum (Range) of vertical , frontal and lateral accelerometer signals. Randomized to 43 subjects for the trained set. We applied stepwise regression to establish the corresponding linear estimation model of walking speed, and used the back-propagation neural network with the same estimation variables to establish walking speed estimation model. Ten subjects for the validation set. The comparison of two models indicates the accuracy of estimation in walking speed. After the statistic analysis, the linear regression equation was obtained by the independent variables (r= 0.93 for trained set, r=0.61 for validation set, RMSE=0.08 m/s for trained set, RMSE=0.242 m/s for validation set). To estimate the walking speed using neural network (r=0.99 for trained set, r=0.98 for validation set, RMSE=0.023 m/s for trained set, RMSE=0.033 m/s for validation set). The relative error for validation set of linear regression and neural network speed estimation is 3.83±1.6%,16.3±0.8%(mean±standard deviation), there is significantly difference (P<0.05). The result shows that using neural networks is more accurate than adopting linear regression analysis in terms of the estimation of a tri-axial accelerometer.
本研究主要以三維加速規應用類神經網路模型來估計步行的平均速度。實驗用的三維加速度規為芬蘭VTI technologies公司所生產的三維加速度晶片(SCA3100-D07),採用取樣頻率200 Hz。利用三軸加速規收集53名受測者(34名男性和19名女性,年齡22.2±1.32歲,身高170.4±7.7公分)行走於50公尺直線走道時的加速度訊號,並紀錄每位受測者實際平均速度,以三個軸向之加速度值方均根、全距值、身高及行走時的步數等八個參數為估測變數,隨機以43名受測者為模型組,應用逐步(stepwise)迴歸分析建立對應的線性步行速度的模型,以相同的估測變數利用類神經網路倒傳遞法建立步行速度的估測模型,10名受測者為驗證組,分別比較兩模型對於估測步行速度的精度。結果顯示線性迴歸所得的無共線性估測變數之步行速度估測模型,對應之模型、驗證組的相關性分別為 r = 0.93、0.61,均方根誤差(RMSE, root mean squared error)分別為0.08、0.242 m/s,類神經網路所得的步行速度估測模型,模型、驗證組的相關性分別為 r = 0.99、0.98,RMSE 分別為0.023、0.033 m/s,兩模型於驗證組的相對誤差分別為 3.83±1.6%、16.3±0.8%,達顯著差異(P < 0.01)。說明了以類神經網路模型用於三維加速規之步行速度估測,顯著優於線性迴歸分析。
文章公開時間: 2017-07-07
Appears in Collections:應用數學系所



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.