這篇論文主要是利用感測器來實現手臂的連續動作辨識。為了達到這個目的，我們使用了IMU並且實作出sensor校正，去除重力，資料平滑化，分割資料以及辨識的方法。我們事先定義了多種的基本手勢。透過這些基本手勢我們得以組合成不同的複合手勢。利用組合的方式也讓我們系統能比較容易的加入新的手勢，讓系統的擴展性更好。實驗說明了我們提出的方法可以有效且準確的辨識出不同的連續手勢。

This thesis proposes a sensing and recognition system for continuous arm gestures. The arm motion is picked up by inertial sensors and processed in stages of calibration, gravity removal, smoothing, segmentation, and recognition. To make it scalable, we define basic gestures in terms of which complex gestures can be defined. Experimental results show our approach to be efficient, accurate, and scalable to a large number of continuous gestures without requiring unnatural pausing.

Contents ……………………………………………………………i
Acknowledgments ……………………………………………………v
1 Introduction ………………………………………………………1
1.1 Motivation………………………………………………………1
1.2 Problem Statement………………………………………………2
1.3 Contribution…………………………………………………………2
1.4 Thesis Organization …………………………………………………2
2 Related Work ………………………………………………………3
2.1 Vision-based Model……………………………………………………3
2.2 Sensor-based Models…………………………………………………4
3 Background Theory ……………………………………………………5
3.1 Sensor Calibration…………………………………………………5
3.1.1 Gyroscope Calibration ……………………………………………5
3.1.2 Accelerometer Calibration ………………………………………5
3.1.3 Magnetometer Calibration ………………………………………7
3.2 The Attitude……………………………………………………………8
3.2.1 Rotations ……………………………………………………………9
3.2.2 Quaternion ………………………………………………………10
3.2.3 AHRS Algorithm ………………………………………………12
3.3 Pattern Recognition ………………………………………………15
3.3.1 SMA Low-Pass Filter……………………………………………15
3.3.2 Pattern Matching …………………………………………………16
4 Technical Approach …………………………………………………18
4.1 CPCP…………………………………………………………………18
4.2 Startup phase …………………………………………………………18
4.2.1 Magnetic Declination ………………………………………………19
4.2.2 Gyroscope Calibration Parameter …………………………………19
4.2.3 Calculate Euler Angle ………………………………………………19
4.2.4 Removing Gravity …………………………………………………20
4.2.5 Preprocessing ………………………………………………………20
4.2.6 Stopping Threshold …………………………………………………21
4.3 Recognition Phase ……………………………………………………21
4.3.1 Stopping Detection …………………………………………………22
4.3.2 Segmentation ………………………………………………………22
4.3.3 Gesture Recognition ………………………………………………23
5 System Architecture and Implementation ………………………………29
5.1 Node Subsystem………………………………………………………29
5.2 The Host and System Architecture ……………………………………30
6 Evaluation ……………………………………………………………32
6.1 Experimental Environment …………………………………………32
6.2 Experimental Data…………………………………………………………32
6.3 Base Gesture ……………………………………………………………………33
6.4 Complex Gestures……………………………………………………34
6.5 Time Complexity ……………………………………………………36
6.6 Limitation………………………………………………………………39
7 Conclusions and Future Work …………………………………………40
7.1 Conclusions………………………………………………………………40
7.2 Future Work ………………………………………………………………41

[1] Renqiang Xie, Xia Sun, Xiang Xia, and Juncheng Cao. Similarity matching-based extensible hand gesture recognition. IEEE Sensors Journal, 15(6):3475–3483, 2015.

[2] Mahmoud Elmezain, Ayoub Al-Hamadi, Jorg Appenrodt, and Bernd Michaelis. A Hidden Markov model-based continuous gesture recognition system for hand motion trajectory. In Proc.19th International Conference on Pattern Recognition, 2008. ICPR 2008., pages 1–4. IEEE, 2008.

[3] Zhipeng Liu, Xiujuan Chai, Zhuang Liu, and Xilin Chen. Continuous gesture recognition with hand-oriented spatiotemporal feature. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 3056–3064, 2017.

[4] JianWu, Lu Sun, and Roozbeh Jafari. A wearable system for recognizing American sign language in real-time using IMU and surface EMG sensors. IEEE Journal of Biomedical and Health Informatics, 20(5):1281–1290, 2016.

[5] Pedro Neto, Dário Pereira, J. Norberto Pires, and A. Paulo Moreira. Real-time and continuous hand gesture spotting: an approach based on artificial neural networks. In Proc. 2013 IEEE International Conference on Robotics and Automation (ICRA), pages 178–183. IEEE, 2013.

[6] Iuri Frosio, Federico Pedersini, and N. Alberto Borghese. Autocalibration of MEMS accelerometers.
IEEE Transactions on Instrumentation and Measurement, 58(6):2034–2041, 2009.

[7] Martin John Baker . Math of quaternion. http://www.euclideanspace.com/maths/geometry/rotations/conversions/quaternionToMatrix/index.htm/.

[8] Sebastian O.H. Madgwick, Andrew J.L. Harrison, and Ravi Vaidyanathan. Estimation of IMU and MARG orientation using a gradient descent algorithm. In Proc. 2011 IEEE International Conference on Rehabilitation Robotics (ICORR), pages 1–7. IEEE, 2011.

[9] yzhajlydy. The calibration of mackdwick using magnetometer. https://blog.csdn.net/nemol1990/article/details/21870197/.