本論文旨在開發一套以慣性元件為基礎之游泳表現分析系統，達到協助運動員自主訓練之效果。系統中之慣性訊號感測器包含了加速度計、陀螺儀、磁力計，用於同步蒐集選手在游泳時背部與頭部所產生之慣性訊號。訊號將透過本論文提出之游泳指標性參數計算演算法計算獲得各項游泳參數。演算法依序擷取游泳區間、偵測轉身與轉身時間、計算單趟划手次數/頻率、評估身體動作一致性、辨識泳姿、計算單趟換氣次數/時間、偵測蹬牆與計算水下打水時間及估測自由式單趟瞬時速度等各項游泳參數。本論文共有25位運動員參與收案(7位大學一般組選手與18位青少年甲組選手)，並以水下影像作為黃金標準，使用留一受試者交叉驗證法(LOSOCV)進行系統驗證。結果顯示，計算單趟划手次數/頻率之準確率可達99.9%；動作一致性評估部分，採用動態時間扭曲演算法(DTW)計算單趟中每次划手時身體的一致性，大學一般組平均一致性為74.44%，青少年甲組一致性為77.58%；在泳姿辨識部分，使用支持向量機器(SVM)進行分類，其平均準確率可達100%；在計算單趟換氣次數/時間之準確率可達99.92%；最後，在自由式單趟瞬時速度變化曲線的計算上，本論文使用適應性權重調變之擴展式卡爾曼濾波器(AWA-based EKF)、座標轉換與重力補償及加速度積分，將背部之慣性訊號轉換為前進方向之速度變化曲線，大學一般組平均速度為0.836(m⁄s)，青少年甲組平均速度為1.116(m⁄s)。研究結果驗證了本系統實際應用於游泳選手訓練的可行性，未來可以進一步的結合雲端資料庫，記錄選手每天訓練的結果，累積成個人之訓練歷程，並藉此調整訓練內容，提升訓練效率，並獲取更好的成績。

This thesis aims to develop an inertial-signal-based swimming performance analysis system (ISPAS) for athletes to conduct self-training. The ISPAS consists of two inertial-sensing-based modules, and each module embeds a triaxial accelerometer, a triaxial gyroscope, and a triaxial magnetometer. A swimming performance analysis (SPA) algorithm has been proposed to analyze the collected inertial signals and provide performance indices for users. This algorithm automatically classifies swimming and non-swimming periods, and detects the following items or indices: swimming types, the number of turns and the time for making a turn, stroke counts and frequency, swimming stroke consistency, breath counts, the time of underwater kicking, and freestyle swimming instantaneous velocity. A total of 25 athletes were recruited for system effectiveness verification. An underwater camera was used to record the swimming activities, and professional coaches analyzed the video to generate the gold standards for the aforementioned items and indices. The results show that, the average accuracy of stroke counts/frequency and breath counts/time can reach 99.9%. For swimming stroke consistency analysis, the average consistency of college students was 74.44%, and the average consistency of teenagers was 77.58%. For the swimming type classification, the average accuracy is up to 100%. Finally, for the freestyle swimming instantaneous velocity, an adaptive weighting adjustment-based extended Kalman filter (AWA-based EKF) were utilized to calculate swimming instantaneous velocity in forward direction. The results validate the effectiveness of the SPA system for providing swimming indices of self-training. In the future, a cloud database could be constructed for athletes to upload their daily training data. Consequently, their coaches can use these data to modify training programs, improve training efficiency, and obtain better performance.

[1] 教育部體育署，泳起來專案，pp. 36，2010。
[2] Y. Ohgi, M. Yasumura, H. Ichikawa, and C. Miyaji. “Analysis of stroke technique using acceleration sensor IC in freestyle swimming,” The engineering of sport, vol. 250, pp. 503-511, 2000.
[3] D. A. James, R. I. Leadbetter, M. R. Neeli, B. J. Burkett, D. V. Thiel, and J. B. Lee. “An integrated swimming monitoring system for the biomechanical analysis of swimming strokes,” Sports Technology, vol. 4, no. 3-4, pp. 141-150, 2011.
[4] F. Dadashi, F. Crettenand, G. P. Millet, L. Seifert, J. Komar, and K. Aminian. “Automatic front-crawl temporal phase detection using adaptive filtering of inertial signals,” Journal of sports sciences, vol. 31, no. 11, pp. 1251-1260, 2013.
[5] N. P. Davey, D. A. James, and M. E. Anderson. “Signal analysis of accelerometry data using gravity-based modeling,” in Proc. of the Microelectronics, MEMS, and Nanotechnology, 2004.
[6] N. Davey, M. Anderson, and D. A. James. “Validation trial of an accelerometer‐based sensor platform for swimming,” Sports Technology, vol. 1, no. 4‐5, pp. 202-207, 2008.
[7] P. Siirtola, P. Laurinen, J. Röning, and H. Kinnunen. “Efficient accelerometer-based swimming exercise tracking,” in Proc. of the 2011 IEEE Symposium on Computational Intelligence and Data Mining (CIDM), 2011
[8] O. Thomas, P. Sunehag, G. Dror, S. Yun, S. Kim, M. Robards, and P. Saunders. “Wearable sensor activity analysis using semi-Markov models with a grammar,” Pervasive and Mobile Computing, vol. 6, no. 3, pp. 342-350, 2010.
[9] M. Bächlin, and G. Tröster. “Swimming performance and technique evaluation with wearable acceleration sensors,” Pervasive and Mobile Computing, vol. 8, no. 1, pp. 68-81, 2012.
[10] T. Le Sage, A. Bindel, P. Conway, L. Justham, S. Slawson, and A. West. “Embedded programming and real-time signal processing of swimming strokes,” Sports Engineering, vol. 14, no. 1, pp. 1, 2011.
[11] B. V. Wright, and J. M. Stager. “Quantifying competitive swim training using accelerometer-based activity monitors,” Sports Engineering, vol. 16, no. 3, pp. 155-164, 2013.
[12] T. Le Sage, P. Conway, S. Slawson, and A. West. “Development of a wireless sensor network for use as an automated system for monitoring swimming starts,” Proceedings of the Institution of Mechanical Engineers, Part P: Journal of Sports Engineering and Technology, vol. 227, no. 3, pp. 184-195, 2013.
[13] F. Dadashi, F. Crettenand, G. P. Millet, and K. Aminian. “Front-crawl instantaneous velocity estimation using a wearable inertial measurement unit,” Sensors, vol. 12, no. 10, pp. 12927-12939, 2012.
[14] T. Le Sage, P. Conway, L. Justham, S. Slawson, A. Bindel, and A. West. “A component based integrated system for signal processing of swimming performance,” in Proc. of the 2010 International Conference on the Signal Processing and Multimedia Applications (SIGMAP), 2010.
[15] T. Le Sage, A. Bindel, P. Conway, L. Justham, S. Slawson, and A. West. “Development of a real time system for monitoring of swimming performance,” Procedia Engineering, vol. 2, no. 2, pp. 2707-2712, 2010.
[16] T. Instruments. CC2640F128RHB datasheet.
[17] STMicroelectronics. LSM9DS1 datasheet.
[18] Macronix. MX25L51245Gz2I datasheet.
[19] S.-h. P. Won, and F. Golnaraghi. “A triaxial accelerometer calibration method using a mathematical model,” IEEE Transactions on Instrumentation and Measurement, vol. 59, no. 8, pp. 2144-2153, 2010.
[20] C. C. Foster, and G. H. Elkaim. “Extension of a two-step calibration methodology to include nonorthogonal sensor axes,” IEEE Transactions on Aerospace and Electronic Systems, vol. 44, no. 3, 2008.
[21] C. Verplaetse. “Inertial proproceptive devices: Self-motion-sensing toys and tools,” IBM Systems Journal, vol. 35, no. 3.4, pp. 639-650, 1996.
[22] U. Jensen, F. Prade, and B. M. Eskofier. “Classification of kinematic swimming data with emphasis on resource consumption,” in Proc. of the 2013 IEEE International Conference on the Body Sensor Networks (BSN), 2013.
[23] 黃思胤，蝶式划手動作技術分析，新竹教育大學體育研究所碩士論文，2013。
[24] 張君賢，動作協調能力之初探，國立臺北教育大學體育研究所碩士論文，2007。
[25] Y.-L. Hsu, C.-L. Chu, Y.-J. Tsai, and J.-S. Wang. “An inertial pen with dynamic time warping recognizer for handwriting and gesture recognition,” IEEE Sensors Journal, vol. 15, no. 1, pp. 154-163, 2015.
[26] D. J. Berndt, and J. Clifford. “Using dynamic time warping to find patterns in time series,” in Proc. of the KDD workshop, 1994.
[27] V. Vapnik. “The nature of statistical learning theory,” Springer science & business media, 2013.
[28] E. Munguia Tapia. “Using machine learning for real-time activity recognition and estimation of energy expenditure,” Massachusetts Institute of Technology, 2008.
[29] A. R. Karduna, P. W. McClure, and L. A. Michener. “Scapular kinematics: effects of altering the Euler angle sequence of rotations,” Journal of Biomechanics, vol. 33, no. 9, pp. 1063-1068, 2000.