本論文主旨在以加速度計實現具身體活動分類與代謝當量估測之演算法，利用配戴於慣用手之手腕、慣用腳之腳踝及腰部之三軸加速度感測器來收集日常活動訊號，先建構活動分類器來辨識靜態類、家事類、樓梯類、步行類及腳踏車類等五種活動類別，再依各活動類別及K4b2氣體分析儀量測之耗氧量來分別建立代謝當量映射迴歸模型以完成準確之熱量估測演算法，而提供使用者有效之活動類別及熱量消耗資訊。
進行活動分類與代謝當量估測演算法前，量測之加速度訊號必須先經過校正、重力濾除與視窗化等前處理程序，再取用前處理後之加速度訊號來擷取三個感測器之count值、腳踝三軸加速度訊號之向量合力的變異係數與腳踝y軸加速度之主頻與次主頻訊號振幅比值等五個特徵值作為活動分類器之輸入，用來訓練機率類神經網路以進行活動分類；而count特徵值也將搭配使用者個人參數(身高、體重、性別及年齡)與實際活動耗氧量，來建構每類活動的代謝當量映射迴歸模型以估測活動之熱量消耗。
本論文結合活動分類器與代謝當量迴歸模型完成一適用於廣泛活動之熱量消耗估測演算法，且活動分類器對於14種日常活動進行驗證後辨識率可達95.33%，而結合分類器所建立之多迴歸模型對於樓梯類及腳踏車類活動之平均估測誤差分別為0.21±1.51 METs與0.16±0.97 METs，相較於分類前以單一迴歸估測模型所產生之-2.05±2.23 METs與0.67±1.24 METs平均估測誤差，其估測準確度獲得明顯改善。

This thesis presents a series of algorithms for physical activity classification and MET mapping model construction using accelerometers. During the measurement of daily activities, three triaxial accelerometer modules are worn at the subject’s dominant wrist, dominant ankle, and waist to collect the accelerations of daily physical activity signals, respectively. The accelerations are used to construct a physical activity classifier, and recognize five activity categories including static, housework, stairs climbing, ambulation, bicycling. Then, the oxygen uptake measured by the Cosmed K4b2 indirect calorimeter for each activity category is used to construct MET mapping regression models.
Before the development of physical activity classifiers and MET mapping model construction algorithm, the first procedure consists of three steps: calibration, gravitational acceleration filtering and windowing. Then, the features extracted from the processed accelerations include counts of three accelerometers, the coefficient of variation from the signal vector magnitude of ankle’s sensor, and the ratio of amplitudes of principal frequency and second frequency from the y-axis signal of ankle’s sensor. The five features are the inputs to train a probability neural network (PNN) to recognize activity categories. Finally, the count value considering the user’s characters such as height, weight, age, and gender, and the oxygen uptake are the features to construct MET mapping regression models to estimate activity energy expenditure.
The thesis combines physical activity classifiers with MET mapping regression models to form an energy expenditure estimation algorithm for a wide range of daily activities. In our experiments, the recognition rate of 14 daily activities using the proposed physical activity classifier is 95.33%, and the errors of average estimation with the multiple regression models of the activity classifier for stairs climbing and bicycling are 0.21±1.51 METs and 0.16±0.97 METs. They are better than the single regression model whose error of stairs climbing and bicycling are -2.05±2.23 METs and 0.67±1.24 METs, respectively. The accuracy of energy expenditure estimation after classification is improved significantly.

[1] I. M. Lee, H. D. Sesso, and R. S. Paffenbarger Jr, “Physical activity and coronary heart disease risk in men does the duration of exercise episodes predict risk,” Circulation, vol. 102, no. 9, pp. 981-986, 2000.
[2] B. A. Calton, R. Z. Stolzenberg-Solomon, S. C. Moore, A. Schatzkin, C. Schairer, D. Albanes, and M. F. Leitzmann, “A prospective study of physical activity and the risk of pancreatic cancer among women (United States),” BioMed Central Cancer, vol. 8, no. 1, pp. 1-9, 2008.
[3] WHO. World Health Report 2002: reducing risks, promoting healthy life. Geneva: World Health Organization, 2002.
[4] 林文元、黃于華，專業就是力量MSN專業健康減重法安全有效，中國醫訊，民國97年。
[5] E. M. Tapia, “Using Machine Learning for Real-time Activity Recognition and Estimation of Energy Expenditure,” Ph.D. dissertation, Massachusetts Institute of Technology, 2008.
[6] M. P. Rothney, E. V. Schaefer, M. M. Neumann, L. Choi, and K. Y. Chen, “Validity of physical activity intensity predictions by Actigraph, Actical, and RT3 accelerometers,” Obesity Society, vol. 16, no. 8, pp. 1946-1952, 2008.
[7] S. E. Crouter, J. R. Churilla, and D. R. Bassett Jr, “Estimating energy expenditure using accelerometers,” Eur. Journal of Applied Physiology, vol. 98, no. 6, pp. 601-612, 2006.
[8] D. R. Bassett Jr, B. E. Ainsworth, A. M. Swartz, S. J. Strath, W. L. O’brien, and G. A. King, “Validity of four motion sensors in measuring moderate intensity physical activity,” Medicine & Science in Sports & Exercise, vol. 32, no. 9, pp. S471-480, 2000.
[9] H. Kumahara, Y. Schutz, M. Ayabe, M. Yoshioka, Y. Yoshitake, M. Shindo, K. Ishii, and H. Tanaka, “The use of uniaxial accelerometry for the assessment of physical-activity-related energy expenditure: a validation study against whole-body indirect calorimetry,” British Journal of Nutrition, vol. 91, no. 2, pp. 235-243, 2004.
[10] S. A. S. Gropper, J. L. Smith, and J. L. Groff, Advanced nutrition and human metabolism. United States of America, Peter Marshall, 2005.
[11] J. A. Harris and F. G. Benedict, “A biometric study of human basal metabolism,” Physiology: Harris and Benedict, vol. 4, no. 12, pp. 370-373, 1918.
[12] A. M. Swartz, S. J. Strath, D. R. Bassett, W. L. O’brien, G. A. King, and B. E. Ainsworth, “Estimation of energy expenditure using CSA accelerometers at hip and wrist sites,” Medicine & Science in Sports & Exercise, vol. 32, no. 9, pp. S450-S456, 2000.
[13] B. Weber, I. Hermanns, R. Ellegast, and J. Kleinert, “A person-centered measurement system for quantification of physical activity and energy expenditure at workplaces,” Ergonomics and Health Aspects, vol. 5624, pp. 121-130, 2009.
[14] S. Schulz, K. R. Westerterp, and K. Brück, “Comparison of energy expenditure by the double labeled water technique with energy intake, heart rate, and activity recording in man,” American Society for Clinical Nutrition, vol. 49, no. 6, pp. 1146-1154, 1989.
[15] G. Plasqui, A. M. C. P. Joosen, A. D. Kester, A. H. C. Goris, and K. R. Westerterp, “Measuring free-living energy expenditure and physical activity with triaxial accelerometer,” Obesity Research, vol. 13, no. 8, pp. 1363-1369, 2005.
[16] S. J. Strath, D. R. Basseett, A. M. Swartz, and D. L. Thompson, “Simultaneous heart rate-motion sensor technique to estimate energy expenditure,” Medicine & Science in Sports & Exercise, vol. 33, no. 12, pp. 2118-2123, 2001.
[17] A. G. Brooks, S. M. Gunn, R. T. Withers, C. J. Gore, and J. L. Plummer, “Predicting walking METs and energy expenditure from speed or accelerometry,” Medicine & Science in Sports & Exercise, vol. 37, no. 7, pp. 1216-1223, 2005.
[18] K. Y. Chen and D. R. Bassett, “The technology of accelerometry-based activity monitors: current and future,” Medicine & Science in Sports & Exercise, vol. 37, no. 11, pp. S490-S500, 2005.
[19] M. P. Rothney, M. Neumann, and A. Béziat, “An artificial neural network model of energy expenditure using nonintegrated acceleration signals,” Eur. Journal of Applied Physiology, vol. 103, no. 4, pp. 1419-1427, 2007.
[20] P. S. Freedson, E. Melanson, and J. Sirard, “Calibration of the computer science and applications, Inc. accelerometer,” Medicine & Science in Sports & Exercise, vol. 30, no. 5, pp. 777-781, 1998.
[21] D. Hendelman, K. C. Baggett, E. Debold, and P. Freedson, “Validity of accelerometry for the assessment of moderate intensity physical activity in the field,” Medicine & Science in Sports & Exercise, vol. 32, no. 9, pp. S442-S449, 2000.
[22] R. S. Rawson and T. M. Walsh, “Estimation of resistance exercise energy expenditure using accelerometry,” Medicine & Science in Sports & Exercise, vol. 42, no. 3, pp. 622-628, 2010.
[23] C. A. Dorminy, L. Choi, S. A. Akohoue, K. Y. Chen, and M. S. Buchowski, “Validity of a multisensor armband in estimating 24h energy expenditure in children,” Medicine & Science in Sports & Exercise, vol. 40, no. 4, pp. 699-706, 2008.
[24] D. R. Bouchard and F. Trudeau, “Estimation of energy expenditure in a work environment: Comparison of accelerometry and oxygen consumption/heart rate regression,” Ergonomics, vol. 51, no. 5, pp. 663-670, 2008.
[25] D. M. Karantonis, M. R. Narayanan, M. Mathie, N. H. Lovell, and B. G. Celler, “Implementation of a real-time human movement classifier using a triaxial accelerometer for ambulatory monitoring,” IEEE Trans. Information Technology in Biomedicine, vol. 10, no. 1, pp. 156-167, 2006.
[26] G. Rodriguez, L. Michaud, L. A. Moreno, D. Turck, and F. Gottrand, “Comparison of the TriTrac-R3D accelerometer and a self-report activity diary with heart-rate monitoring for the assessment of energy expenditure in children,” British Journal of Nutrition, vol. 87, no. 6, pp. 623-631, 2002.
[27] G. J. Welk, S. N. Blair, K. Wood, S. Jones, and R. W. Thompson, “A comparative evaluation of three accelerometry-based physical activity monitors,” Medicine & Science in Sports & Exercise, vol. 32, no. 9, pp. S489-S497, 2000.
[28] D. M. Pober, J. Staudenmayer, C. Raphael, and P. S. Freedson, “Development of novel techniques to classify physical activity mode using accelerometers,” Medicine & Science in Sports & Exercise, vol. 38, no. 9, pp. 1626-1634, 2006.
[29] J. A. Toschke, R. V. Kries, and E. Rosenfeld, “Reliability of physical activity measures from accelerometry among preschoolers in free-living conditions,” Clinical Nutrition, vol. 26, no. 4, pp. 416-420, 2007.
[30] S. E. Crouter, and D. R. Bassett Jr, “A new 2-regression model for the Actical accelerometer,” British Journal of Sports Medicine, vol. 42, no. 3, pp. 217-224, 2008.
[31] C. V. C. Bouten, K. T. M. Koekkoek, M. Verduin, R. Kodde, and J. D. Janssen, “A triaxial accelerometer and portable data processing unit for assessment the assessment of daily physical activity,” IEEE Trans. Biomedical Engineering, vol. 44, no. 3, pp. 136-147, 1997.
[32] W. T. Ang, P. K. Khosla, and C. N. Riviere, “Nonlinear regression model of a low-g MEMS accelerometer,” IEEE Sensors Journal, vol. 7, no. 1, pp. 81-88, 2007.
[33] I. Frosio, F. Pedersini, and N. A. Borghese, “Autocalibration of MEMS accelerometers,” IEEE Trans. Instrumentation and Measurement, vol. 58, no. 6, 2009.
[34] 楊諄縈，使用單一加速度計及特徵降維之類神經辨識器於人類動作辨識，國立成功大學電機系碩士論文，2008。
[35] S. J. Preece, J. Y. Goulermas, L. P. J. Kenney, D. Howard, K. Meijer, and R. Crompton,“Activity identification using body-mounted sensors-a review of classification techniques,” Physiological Measurement, vol. 30, no. 4, pp. R1-R33, 2009.
[36] S. E. Crouter, K. G. Clowers, and D. R. Bassett Jr, “A novel method for using accelerometer data to predict energy expenditure,” Eur. Journal of Applied Physiology, vol. 100, no. 4, pp. 1324-1331, 2006.
[37] Y. Wang, L. Li, J. Ni, and S. Huang, “Feature selection using tabu search with long-term memories and probabilistic neural networks,” Pattern Recognition Letters, vol. 30, no. 7, pp. 661-670, 2009.
[38] 行政院衛生署，每日營養素建議攝取量及其說明，民國82年。
[39] E. K. P. Chong and S. H. An introduction to optimization. Hoboken, NJ: John Wiley and Sons, 2008.
[40] 戴久永，統計概念與方法，民國88年8月。