近年來國內健身風氣越來越盛行，不管是為了修飾體態或是健體強身，重量訓練都是運動計畫中相當重要的一環，無奈國內外疫情肆虐，居家健身成了人們運動的主要形式，啞鈴的便利與功能性則是居家重量訓練器材的首選。拜科技進步所賜，科學化運動除了運用在職業比賽上，也漸漸被帶入市井小民的手中，智慧化的紀錄運動數據能夠幫助訓練者檢視自己的運動歷程，進而提升訓練成效與運動意願。
常見的運動紀錄系統包含兩個主要功能，動作計次以及動作辨識，為達成此兩種功能，本論文提出以藍牙穿戴式裝置收集慣性感測器(IMU)訊號與藍牙接收訊號強度(RSSI)來進行啞鈴的動作計次和動作辨識。其中動作計次是利用慣性感測器訊號作為判斷依據，藉由座標轉換與波峰波谷偵測來偵測運動週期。動作辨識則額外引入了RSSI訊號，透過隨機森林排序RSSI與IMU特徵的價值，選出最有效的特徵集，且在多種分類器中評估出隨機森林分類器為效能最佳分類器。
本論文提出之系統，實驗結果顯示即時動作計次的平均準確度為99.5%，即時動作辨識的平均準確度為96.7%，且實驗結果證實RSSI能有效提高IMU所無法辨識的動作種類的準確度。

Weight training has multiple health benefits such as improving strength and endurance and reducing health risks. During the COVID-19 pandemic, exercising at home has become the main trend for people all over the world. The convenience and functionality make dumbbells become the first choice for home weight training equipment. Without the assistance of the coach, automatic recording of exercise results by an exercise monitoring system can help users review their own exercise history, thereby improving training effectiveness and exercise willingness.
In this thesis, we proposed a dumbbell exercise repetition counting and recognition system based on Bluetooth wearable devices, which acquire motion signals from inertial measurement units (IMU) and Bluetooth Received Signal Strength Indication (RSSI). The exercise repetition counting algorithm is based on the coordinate conversion and peak and trough detection of IMU signals. Activity recognition additionally introduces RSSI signals, the importance of IMU and RSSI feature sets are sorted by Random Forest first and evaluated with four types of classifiers which include SVM, KNN, Naive Bayes and Random Forest, the Random Forest classifier in combination with RSSI feature set come out the best performance.
The experimental results of the system proposed in this thesis show that the average accuracy of real-time repetition counting is 99.5%, and the average accuracy of real-time activity recognition is 96.7%. Besides, the results also show that the employment of RSSI can effectively improve the recognition accuracy of the types of exercises that the IMU-based system cannot recognize.

[1] i運動資訊平台-教育部體育署, "中華民國109年運動現況調查結案報告書," 2020.
[2] W. H. O., Global Recommendations on Physical Activity for Health. World Health Organization, 2010.
[3] Freeletics. (2020). Freeletics surveys Americans to uncover what is in store for the fitness industry post-COVID-19. Available: https://www.freeletics.com/en/press/news/freeletics-surveys-americans-to-understand-what-is-in-store-for-the-fitness-industry-post-covid-19/
[4] A. Nagarkoti, R. Teotia, A. K. Mahale, and P. K. Das, "Realtime Indoor Workout Analysis Using Machine Learning & Computer Vision," in 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2019, pp. 1440-1443.
[5] G. Dsouza, D. Maurya, and A. Patel, "Smart Gym Trainer Using Human Pose Estimation," in 2020 IEEE International Conference for Innovation in Technology (INOCON), 2020, pp. 1-4.
[6] H. Ding et al., "A Platform for Free-Weight Exercise Monitoring with Passive Tags," IEEE Transactions on Mobile Computing, vol. 16, no. 12, pp. 3279-3293, 2017.
[7] X. Guo, J. Liu, C. Shi, H. Liu, Y. Chen, and M. C. Chuah, "Device-free Personalized Fitness Assistant Using WiFi," Proc. ACM Interact. Mob. Wearable Ubiquitous Technol., vol. 2, pp. 1-23, 2018.
[8] P. Hausberger, A. Fernbach, and W. Kastner, "IMU-based Smart Fitness Devices for Weight Training," in IECON 2016 - 42nd Annual Conference of the IEEE Industrial Electronics Society, 2016, pp. 5182-5189.
[9] X. Guo, J. Liu, and Y. Chen, "FitCoach: Virtual Fitness Coach Empowered by Wearable Mobile Devices," in IEEE INFOCOM 2017 - IEEE Conference on Computer Communications, 2017, pp. 1-9.
[10] J. Qi, P. Yang, M. Hanneghan, S. Tang, and B. Zhou, "A Hybrid Hierarchical Framework for Gym Physical Activity Recognition and Measurement Using Wearable Sensors," IEEE Internet of Things Journal, vol. 6, no. 2, pp. 1384-1393, 2019.
[11] Y. Zou, D. Wang, S. Hong, R. Ruby, D. Zhang, and K. Wu, "A Low-Cost Smart Glove System for Real-Time Fitness Coaching," IEEE Internet of Things Journal, vol. 7, no. 8, pp. 7377-7391, 2020.
[12] 高雅珩, "基於機器學習的即時啞鈴運動識別系統," 碩士論文, 國立臺灣科技大學, 2020.
[13] A. Filippeschi, N. Schmitz, M. Miezal, G. Bleser, E. Ruffaldi, and D. Stricker, "Survey of Motion Tracking Methods Based on Inertial Sensors: A Focus on Upper Limb Human Motion," (in eng), Sensors (Besel, Switzerland), vol. 17, no. 6, 2017.
[14] M. Euston, P. Coote, R. Mahony, J. Kim, and T. Hamel, "A Complementary Filter for Attitude Estimation of a Fixed-wing UAV," in 2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, 2008, pp. 340-345.
[15] 鄭期元, "整合GNSS與INS量測資訊的地平面運動軌跡估測," 碩士論文, 電控工程研究所, 國立交通大學, 新竹市, 2013.
[16] W. R. Hamilton, "On Quaternions; or On a New System of Imaginaries in Algebra," Philosophical magazine, vol. 25, no. 3, pp. 489-495, 1844.
[17] J. B. Kuipers, Quaternions and Rotation Sequences: a Primer with Applications to Orbits, Aerospace, and Virtual Reality. Princeton university press, 1999.
[18] V. Gao. (2015, 2021, July 5). Proximity and RSSI. Available: https://www.bluetooth.com/blog/proximity-and-rssi/
[19] F. Itakura, "Minimum Prediction Residual Principle Applied to Speech Recognition," IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. 23, no. 1, pp. 67-72, 1975.
[20] C. Shyalika. (2019). Dynamic Time Warping Algorithm for Time Series Analysis. Available: https://medium.datadriveninvestor.com/dynamic-time-warping-dtw-d51d1a1e4afc
[21] S. Salvador and P. Chan, "Toward Accurate Dynamic Time Warping in Linear Time and Space," Intelligent Data Analysis, vol. 11, no. 5, pp. 561-580, 2007.
[22] C. Cortes and V. Vapnik, "Support-vector networks," Machine learning, vol. 20, no. 3, pp. 273-297, 1995.
[23] L. Breiman, "Random Forests," Machine Learning, vol. 45, no. 1, pp. 5-32, 2001 2001.
[24] H. Zhang, "The Optimality of Naive Bayes," presented at the Proceedings of the Seventeenth International Florida Artificial Intelligence Research Society Conference, 2004.
[25] N. S. Altman, "An Introduction to Kernel and Nearest-Neighbor Nonparametric Regression," The American Statistician, vol. 46, no. 3, pp. 175-185, 1992.
[26] Sam. Arm Workout. Available: https://www.pinterest.com/sew1090439/arm-workout/
[27] N. Semiconductor. (2017). nRF52832 Product Specification v1.4. Available: https://infocenter.nordicsemi.com/index.jsp?topic=%2Fstruct_nrf52%2Fstruct%2Fnrf52832_ps.html
[28] InvenSense. (2016). ICM-20649. Available: https://product.tdk.com/info/en/documents/catalog_datasheet/imu/DS-000192-ICM-20649-v1.0.pdf
[29] F. Pedregosa et al., "Scikit-learn: Machine Learning in Python," Journal of Machine Learning Research, vol. 12, pp. 2825--2830, 2011.
[30] S. Shen, M. Gowda, and R. R. Choudhury, "Closing the Gaps in Inertial Motion Tracking," presented at the Proceedings of the 24th Annual International Conference on Mobile Computing and Networking, 2018.
[31] R. Chen, C. Dewi, S. Huang, and R. E. Caraka, "Selecting Critical Features for Data Classification Based on Machine Learning Methods," Journal of Big Data, vol. 7, no. 1, 2020.