隨著高齡化社會的發展，市面上的外骨骼產品越來越多元，其中助力型外骨骼不僅可以減少勞動負擔，也可以避免勞動損傷的情況發生，醫療型外骨骼可以提供傷患便利的復健過程。另一方面，此類型的外骨骼裝置，在發生步態切換時，外骨骼無法達到即時的控制模式切換，容易造成不舒適的頓挫感。在此背景下，本文提出一套穩定步態與步態轉換辨識架構，採用深度學習預訓練模型，能夠準確辨識出不同的步態切換情況發生。在本文的系統架構中，各感測器訊號由Raspberry Pi 接收，分別收取訓練集和測試集資料，透過訓練集資料建立預訓練模型，接著使用預訓練模型對測試集資料做分析，即可透過離線的方式，分析各個步態轉換情況下，預訓練模型的分類辨識效果。在系統驗證方面，針對四種生活中常見的步態，以及各步態間可能發生之轉換區間，個別收取一定筆數之測試集資料，針對模型辨識之準確率及提前時間等指標進行分析，針對各步態轉換區間都能有良好的辨識效果，並且在正式進入下一個穩定步態之前，能夠提供一段緩衝時間。實驗結果證明了本論文開發之系統在實務上的適用性，對於步態轉換之分類辨識具有良好的參考價值。

With the development of an aging society, there is an increasing variety of exoskeleton products available on the market. Among them, assistive exoskeletons not only reduce physical burdens but also help prevent work-related injuries. Medical exoskeletons facilitate the rehabilitation process for patients. On the other hand, this type of exoskeleton device lacks real-time control mode switching during gait transitions, resulting in uncomfortable jolts. In this context, this paper proposes a stable gait and gait transition recognition framework, which utilizes a pre-trained convolutional neural network model to accurately identify different gait transition interval. In the system architecture presented in this paper, sensor signals are received by a Raspberry Pi, which collects training and testing data. The pre-trained model is built using the training data, and subsequently, the testing data is analyzed using the pre-trained model. This offline analysis allows for the evaluation of the classification performance of the pre-trained model for various gait transition interval. In terms of system verification, a certain number of testing data sets are collected for each of the four common gaits and possible transition intervals between gaits. The accuracy and lead time of the model's recognition are analyzed as performance indicators. The results demonstrate that the proposed system in this paper has practical applicability, providing good recognition accuracy for gait transitions and a buffer period before entering the next stable gait. The experimental results validate the practicality of the developed system and its valuable reference for gait transition classification and recognition.

[1] 黃志仁。「穿戴式下肢輔具之研發設計及效能評估」。博士論文，國立清華大學動力機械工程學系，2011。＜https://hdl.handle.net/11296/q3hrzx＞。
[2] H. Kawamoto and Y. Sankai, "Comfortable power assist control method for walking aid by HAL-3," IEEE International Conference on Systems, Man and Cybernetics, Yasmine Hammamet, Tunisia, 2002, pp. 6 pp. vol.4-, doi: 10.1109/ICSMC.2002.1173328.
[3] K. Seo, J. Lee, Y. Lee, T. Ha and Y. Shim, "Fully autonomous hip exoskeleton saves metabolic cost of walking," 2016 IEEE International Conference on Robotics and Automation (ICRA), Stockholm, Sweden, 2016, pp. 4628-4635, doi: 10.1109/ICRA.2016.7487663.
[4] K. Seo, J. Lee and Y. J. Park, "Autonomous hip exoskeleton saves metabolic cost of walking uphill," 2017 International Conference on Rehabilitation Robotics (ICORR), London, UK, 2017, pp. 246-251, doi: 10.1109/ICORR.2017.8009254.
[5] Y. Ding, I. Galiana, C. Siviy, F. A. Panizzolo and C. Walsh, "IMU-based iterative control for hip extension assistance with a soft exosuit," 2016 IEEE International Conference on Robotics and Automation (ICRA), Stockholm, Sweden, 2016, pp. 3501-3508, doi: 10.1109/ICRA.2016.7487530.
[6] R. Analia, J. Ferdinand M., Susanto, P. D. Sutopo, H. Soebhakti and A. Sani, "Walking Classification of Hip Joint Lower Limb Exoskeleton," 2019 2nd International Conference on Applied Engineering (ICAE), Batam, Indonesia, 2019, pp. 1-5, doi: 10.1109/ICAE47758.2019.9221701.
[7] Y. Zhang, Z. Ma, S. Zuo and J. Liu, "Gait Prediction and Assist Control of Lower Limb Exoskeleton Based on Inertia Measurement Unit," 2022 5th International Conference on Intelligent Robotics and Control Engineering (IRCE), Tianjin, China, 2022, pp. 44-49, doi: 10.1109/IRCE55557.2022.9963096.
[8] D. Gouwanda, A. A. Gopalai and B. H. Khoo, "A Low Cost Alternative to Monitor Human Gait Temporal Parameters–Wearable Wireless Gyroscope," IEEE Sensors Journal, vol. 16, no. 24, pp. 9029-9035, 15 Dec.15, 2016, doi: 10.1109/JSEN.2016.2616163.
[9] S. Ding, X. Ouyang, T. Liu, Z. Li and H. Yang, "Gait Event Detection of a Lower Extremity Exoskeleton Robot by an Intelligent IMU," IEEE Sensors Journal, vol. 18, no. 23, pp. 9728-9735, 1 Dec.1, 2018, doi: 10.1109/JSEN.2018.2871328.

[10] C. Livolsi, R. Conti, F. Giovacchini, N. Vitiello and S. Crea, "A Novel Wavelet-Based Gait Segmentation Method for a Portable Hip Exoskeleton," IEEE Transactions on Robotics, vol. 38, no. 3, pp. 1503-1517, June 2022, doi: 10.1109/TRO.2021.3122975.
[11] Pirker, W., Katzenschlager, R. Gait disorders in adults and the elderly. Wien Klin Wochenschr 129, 81–95 (2017).
[12] A.C. Novak, S.M. Reid, P.A. Costigan, B. Brouwer. Stair negotiation alters stability in older adults. Lower Extrem. Rev., 2 (10) (2010), pp. 47-51.
[13] K. Xia, J. Huang and H. Wang, "LSTM-CNN Architecture for Human Activity Recognition," IEEE Access, vol. 8, pp. 56855-56866, 2020, doi: 10.1109/ACCESS.2020.2982225.
[14] K. Struebig, N. Ganter, L. Freiberg and T. C. Lueth, "Stair and Ramp Recognition for Powered Lower Limb Exoskeletons," 2021 IEEE International Conference on Robotics and Biomimetics (ROBIO), Sanya, China, 2021, pp. 1270-1276, doi: 10.1109/ROBIO54168.2021.9739447.
[15] LECUN, Yann, et al. "Backpropagation applied to handwritten zip code recognition." Neural computation, 1989, 1.4: 541-551.
[16] 麥峻嘉, "基於MEMS感測器與霧運算平台之智慧物聯網應用於感應電動機狀態監控與故障診斷" 碩士, 電機工程學系, 國立成功大學, 台南市, 2022.
[17] F. Xu, X. Lin, H. Cheng, R. Huang and Q. Chen, "Adaptive stair-ascending and stair-descending strategies for powered lower limb exoskeleton," 2017 IEEE International Conference on Mechatronics and Automation (ICMA), Takamatsu, Japan, 2017, pp. 1579-1584, doi: 10.1109/ICMA.2017.8016052.
[18] J. Jang, K. Kim, J. Lee, B. Lim, J. -K. Cho and Y. Shim, "Preliminary study of online gait recognizer for lower limb exoskeletons," 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Vancouver, BC, Canada, 2017, pp. 5818-5824.
[19] Andrew H. Hansen, Dudley S. Childress, Margrit R. Meier, A simple method for determination of gait events, Journal of Biomechanics, Volume 35, Issue 1, 2002, Pages 135-138, ISSN 0021-9290.
[20] T. Zhang, M. Tran and H. Huang, "Design and Experimental Verification of Hip Exoskeleton With Balance Capacities for Walking Assistance," IEEE/ASME Transactions on Mechatronics, vol. 23, no. 1, pp. 274-285, Feb. 2018, doi: 10.1109/TMECH.2018.2790358.
[21] Q. Ma, L. Ji and R. Wang, "The Development and Preliminary Test of a Powered Alternately Walking Exoskeleton With the Wheeled Foot for Paraplegic Patients," in IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 26, no. 2, pp. 451-459, Feb. 2018, doi: 10.1109/TNSRE.2017.2774295.
[22] T. Lenzi, M. C. Carrozza and S. K. Agrawal, "Powered Hip Exoskeletons Can Reduce the User's Hip and Ankle Muscle Activations During Walking," IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 21, no. 6, pp. 938-948, Nov. 2013, doi: 10.1109/TNSRE.2013.2248749.
[23] W. Yang, L. H. Xu and C. J. Yang, "Adaptive torque control of hip exoskeleton for walking assist based on motion comparison method and phase oscillator method," 2019 Intelligent Rehabilitation and Human-machine Engineering Conference (IRHE-2019), Qinhuangdao, China, 2019, pp. 1-6, doi: 10.1049/cp.2019.1192.
[24] A. J. Young and D. P. Ferris, "State of the Art and Future Directions for Lower Limb Robotic Exoskeletons," IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 25, no. 2, pp. 171-182, Feb. 2017, doi: 10.1109/TNSRE.2016.2521160.
[25] W. Huo, S. Mohammed, J. C. Moreno and Y. Amirat, "Lower Limb Wearable Robots for Assistance and Rehabilitation: A State of the Art," IEEE Systems Journal, vol. 10, no. 3, pp. 1068-1081, Sept. 2016, doi: 10.1109/JSYST.2014.2351491.
[26] P. Li, T. Xue, T. Zhang, H. Wang and M. Zhang, "Hybrid Gait Recognition Method Based on Phase Trajectory," 2022 27th International Conference on Automation and Computing (ICAC), Bristol, United Kingdom, 2022, pp. 1-6, doi: 10.1109/ICAC55051.2022.9911060.
[27] Yang Han, Chunbao Liu, Lingyun Yan and Lei Ren, "Design of Decision Tree Structure with Improved BPNN Nodes for High-Accuracy Locomotion Mode Recognition Using a Single IMU", Sensors, vol. 21, pp. 526, 2021.
[28] I. Kang, D. D. Molinaro, G. Choi and A. J. Young, "Continuous locomotion mode classification using a robotic hip exoskeleton," 2020 8th IEEE RAS/EMBS International Conference for Biomedical Robotics and Biomechatronics (BioRob), New York, NY, USA, 2020, pp. 376-381, doi: 10.1109/BioRob49111.2020.9224359.
[29] I. Kang, P. Kunapuli and A. J. Young, "Real-Time Neural Network-Based Gait Phase Estimation Using a Robotic Hip Exoskeleton," IEEE Transactions on Medical Robotics and Bionics, vol. 2, no. 1, pp. 28-37, Feb. 2020, doi: 10.1109/TMRB.2019.2961749.
[30] Jindong Wang, Yiqiang Chen, Shuji Hao, Xiaohui Peng, Lisha Hu, Deep learning for sensor-based activity recognition: A survey, Pattern Recognition Letters, Volume 119, 2019, Pages 3-11, ISSN 0167-8655.