在本論文中設計出一款低功耗多功能的穿戴式生理感測器，使用五種不同的感測器來量測六種生理訊號，包含穿戴者的心電圖 (Electrocardiography) 訊號、呼吸時的胸腔阻抗變化、三軸運動軌跡、九軸運動軌跡、不同環境的溫濕度變化。並使用微控制器整合感測到的生理訊號，透過藍芽低功耗 (Bluetooth low energy, BLE)將資料傳輸至行動裝置，或者將資料傳送至快閃記憶體 (NAND flash)內儲存。

在本穿戴式感測器的系統設計中，我們提供兩種操作模式: ECG-9axis模式與ECG-3axis模式，讓穿戴者可已依據自身需求切換模式。另外資料儲存方面提供兩種操作版本: BLE版本與NAND flash版本，讓穿戴者在無法使用藍芽低功耗的環境下，也能將資料內存在裝置內。在藍芽低功耗版本下，我們設計藍芽自動回連機制，在藍芽意外斷線的情境下，也能自動與接收裝置回連。另外我們開發了心率偵測演算法，評估心電圖訊號品質。當心電圖訊號品質不佳時，感測器會發出警示燈，提醒穿戴者重新穿戴感測器。

本穿戴式感測器使用一顆300毫安時的電池作為電源供應。在功率消耗方面，未實現系統設計之前，ECG-9axis模式下的平均電流為7.91毫安，ECG-3axis模式下的平均電流為4.75毫安。經過系統設計後，ECG-9axis模式NAND flash版本下平均電流為6.14毫安，減少了22.43%的電流消耗，使用時間約為48.84小時。ECG-3axis模式NAND flash版本下平均電流為3.26毫安，減少了31.39%的電流消耗，使用時間約為91.87小時。

為了延長快閃記憶體的儲存時間，我們實現了無損心電圖訊號壓縮，壓縮率為2.93。在ECG-9axis模式下，儲存時間延長了1.64倍，可儲存45.76小時的資料。在ECG-3axis模式下，儲存時間延長了2.24倍，可儲存85.76小時的資料，近乎是三天半的時間。

除了穿戴式感測器的設計之外，我們也應用感測器內的三軸加速度計，替長期臥床的病患，開發翻身姿態辨識系統，在十一位志願者與一個假人病患的實驗下，48 次的翻身均可百分之百辨識。由於長期臥床的病患需要定時的去翻身避免褥瘡的發生，本系統可以記錄穿戴者翻身的姿態跟時間，另外也可以透過行動裝置上的應用程式監控翻身姿態次數，確保病患的醫療品質與改善人力管理。

In this paper, a low-power multi-functional wearable sensor is designed, which uses five different sensors to measure six physiological signals, including the electrocardiography (ECG) signal, respiration, 3-axis movement trajectory, 9-axis movement trajectory, temperature and humidity in different environment. Meanwhile, microcontrollers are used to integrate the physiological signals detected and transfer data to mobile devices through Bluetooth low energy (BLE) or to NAND flash memory for storage.

In the system design of this wearable sensor, we provide two operation modes: ECG-9axis mode and ECG-3axis mode, allowing the wearer to switch modes according to their own needs. Furthermore, there are two operating versions of data storage: BLE version and NAND flash version, which allow the wearer to store data in the device when the wearer cannot use BLE in some environment. In the BLE version, we designed an automatic reconnection mechanism of BLE, which can also automatically reconnect with the receiver in the case of accidental disconnection of BLE. In addition, we developed a heart rate detection algorithm to evaluate the ECG signal quality. When the ECG signal quality is poor, the sensor will send a warning light to remind the wearer to re-wear the sensor.

We use a 300 mAh battery as power source of the wearable sensor. In terms of power consumption, before the system design was implemented, the average current in ECG-9axis mode was 7.91 mA, and the average current in ECG-3axis mode was 4.75 mA. After the system design, the average current in ECG-9axis mode NAND flash version is 6.14 mA, reducing the current consumption by 22.43\% and the battery lifetime is about 48.84 hours. The average current in ECG-3axis mode NAND flash version is 3.26 mA, reducing current consumption by 31.39\% and the battery lifetime is about 91.87 hours.

In order to extend the storage time of NAND flash, we implemented the lossless compression of ECG signal with a compression rate (CR) is 2.93. In ECG-9axis mode, the storage time is increased by 1.64 times and 45.76 hours of data can be stored. In ECG-3axis, storage time is increased by 2.24 times and 85.76 hours of data can be stored, almost three and a half days.

In addition to the design of wearable sensor, we also use 3-axis accelerometer inside the sensor to develop a turn over posture recognition system for patients who are in bed for a long time. Under the experiment of 11 volunteers and one dummy patient, 48 times turn over could be recognized 100\%. Since patients in bed for a long time need to turn over regularly to avoid bedsores, the system can record the posture and time of the wearer turning over. Beside, the application program on the mobile device can also monitor the times of turning over to ensure the medical quality of patients and improve human resource management.

1 Introduction 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Main Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Features of the Physiological Signals and Wireless Transmission Technologies 5
2.1 Physiological Signals in System . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.1 Electrocardiography (ECG) . . . . . . . . . . . . . . . . . . . . . . 5
2.1.2 Respiration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.3 Motion Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Wireless Transmission Technologies . . . . . . . . . . . . . . . . . . . . . . 13
2.2.1 Bluetooth Low Energy (BLE) . . . . . . . . . . . . . . . . . . . . . 14
2.2.2 BLE Protocol Stack . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2.3 BLE Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 Related Studies Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3 Proposed Physiological Sensing System 21
3.1 System Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2 Front-end Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2.1 ECG and Respiration Sensor (ADS1292R) . . . . . . . . . . . . . . 25
3.2.2 3-axis Sensor (BMA253) . . . . . . . . . . . . . . . . . . . . . . . . 30
3.2.3 9-axis Sensor (BHI160 and BMM150) . . . . . . . . . . . . . . . . . 32
3.2.4 Temperature and Humidity Sensor (HDC2010) . . . . . . . . . . . . 34
3.3 Control-end and Wireless Interface . . . . . . . . . . . . . . . . . . . . . . . 36
3.3.1 Microcontroller (MSP4321P401R) . . . . . . . . . . . . . . . . . . . 36
3.3.2 Wireless Interface (CC2640R2F) . . . . . . . . . . . . . . . . . . . . 39
3.4 Storage Component . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.5 Firmware Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.5.1 Heart Rate Detection . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.5.2 Lossless Compression of ECG Signal . . . . . . . . . . . . . . . . . 45
3.6 Robustness Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.6.1 Bluetooth Reconnection Mechanism . . . . . . . . . . . . . . . . . . 52
3.6.2 Physiological Signal Storage . . . . . . . . . . . . . . . . . . . . . . 53
3.6.3 ECG Quality Indices . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.7 System Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.8 Turn-Over Posture Recognition . . . . . . . . . . . . . . . . . . . . . . . . . 60
4 Implementation Results 65
4.1 Sensor Node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.1.1 Appearance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.1.2 Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.1.3 Physiological Signal Measurement . . . . . . . . . . . . . . . . . . . 68
4.1.4 Data Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.1.5 Power Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.2 Heart Rate Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.3 Lossless Compression of ECG Signal . . . . . . . . . . . . . . . . . . . . . 89
4.4 Turn-Over Posture Recognition . . . . . . . . . . . . . . . . . . . . . . . . . 91
5 Conclusions and Future Works 97
5.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
5.2 Future Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Bibliography 106

[1] M. of the Interior. (2019) Taiwan’s population aging index. Department of Statistics. [Online]. Available: https://www.moi.gov.tw/stat/chart.aspx?ChartID=S0401
[2] M. of Health and Welfare. (2017) Report of the senior citizen condition survey 2017. Department of Statistics. [Online]. Available: https://dep.mohw.gov.tw/dos/cp-1767-38429-113.html
[3] D. Cao, Z. Zhang, H. Liang, X. Liu, Y. She, Y. Li, D. Li, and M. Yu, “Application of a wearable physiological monitoring system in pulmonary respiratory rehabilitation research,” Oct 2018, pp. 1–6.
[4] (2019) Garmin vvosport. Garmin. [Online]. Available: https://www.garmin.com.tw/products/intosports/vivosport-slate-s/
[5] (2019) Sony smartband 2. Sony. [Online]. Available: https://www.sonymobile.com/tw/products/smart-products/smartband-2/
[6] (2019) Mi band 3. Mi. [Online]. Available: https://www.mi.com/tw/mi-band-3/
[7] M. Magno, M. Pritz, P. Mayer, and L. Benini, “Deepemote: Towards multi-layer neural networks in a low power wearable multi-sensors bracelet,” in 2017 7th IEEE International Workshop, Advances in Sensors and Interfaces (IWASI), 2017, pp. 32–37.
[8] Y.Wang, S. Doleschel, R.Wunderlich, and S. Heinen, “A wearable wireless ECG monitoring system with dynamic transmission power control for long-term homecare,” Journal of medical systems, vol. 39, p. 223, Mar 2015.
[9] J. Malmivuo and R. Plonsey, Bioelectromagnetism: principles and applications of bioelectric and biomagnetic fields. Oxford University Press, USA, 1995.
[10] A. S. Ary L. Goldberger, Zachary D. Goldberger, Goldbergers Clinical Electrocardiography: A Simplified Approach. Elsevier Science Health Science, 2017.
[11] S. B. H. L. B. M. W. Kim E. Barrett, Susan M. Barman, Ganong’s review of medical physiology. McGraw Hill Education, 2016.
[12] O. College, Anatomy physiology. OpenStax CNX, 2013.
[13] A. Fedotov, S. Akulov, and A. Akulova, “Motion artifacts reduction in wearable respiratory monitoring device,” Jun 2017, pp. 1121–1124.
[14] J. Lee, J. P Florian, and h. ki, “Respiratory rate extraction from pulse oximeter and electrocardiographic recordings,” Physiological measurement, vol. 32, pp. 1763–73, Nov 2011.
[15] J. Mann, R. Rabinovich, A. Bates, S. Giavedoni, W. MacNee, and D. Arvind, “Simultaneous activity and respiratory monitoring using an accelerometer,” Jun 2011, pp. 139 – 143.
[16] Respiration Rate Measurement Based on Impedance Pneumography, Texas Instruments, 2011.
[17] R. Oreilly, A. Khenkin, and K. Harney, “Sonic nirvana: Using mems accelerometers as acoustic pickups in musical instruments,” vol. 26, Mar 2009.
[18] M. Andrejaic. (2008) MEMS accelerometers. University of Ljubljana Tech. [Online]. Available: http://mafija.fmf.uni Ij.si/seminar/files/2007 2008/MEMS accelerometers-koncna.pdf
[19] J. Geen and D. Krakauer, “New iMEMS R angular-rate-sensing gyroscope,” Analog Dialogue, vol. 37, no. 3, pp. 1–4, 2003.
[20] Measuring magnetic fields. The field of Nondestructive Testing. [Online]. Available: https://www.ndeed.org/EducationResources/CommunityCollege/MagParticle/Physics/Measuring.htm
[21] Sagawa. (2013, Sep) Quick thoughts: For payments, Bluetooth puts a beat down on NFC and WiFis. [Online]. Available: http://ssrllc.com/quick-thoughts-for-payments-bluetooth-puts-a-beatdown-on-nfc-and-wifi/
[22] (2016, Nov) ZigBee in home automation. CCTV Institude. [Online]. Available: http://cctvinstitute.co.uk/zigbee/
[23] CC2640R2F SimpleLink Bluetooth low energy Wireless MCU, Texas Instruments, Dec 2016, [Revised Jan 2017].
[24] Bluetooth 5 Go Faster. Go Futher., Bluetooth SIG, 2018.
[25] CC2640 and CC2650 SimpleLink Bluetooth low energy Software Stack 2.2.1 Developer’s Guide, Texas Instruments, Oct 2016.
[26] G. Zanichelli, “A wearable multi-sensor system for medical diagnostics and patient monitoring,” Master’s thesis, Polytechnic University of Turin, Turin, Italy, 2018.
[27] C. Y. Chen, “An ultra low-power wearable sensor for physiological signal monitoring,” Master’s thesis, National Tsing Hua University, Hsinchu, Taiwan, 2017.
[28] P. C. Hsu, “A configurable robust low-power wearable sensing system,” Master’s thesis, National Tsing Hua University, Hsinchu, Taiwan, 2018.
[29] Low-Power, 2-Channel, 24-Bit Analog Front-End for Biopotential Measurements, Texas Instruments, Dec. 2011, [Revised Sep 2012].
[30] BMA253 Digital, triaxial acceleration sensor, Bosch Sensortec, Aug. 2015, rev. 1.0.
[31] BHI160 Ultra-low power sensor hub incl. integrated IMU, Bosch Sensortec, Jul. 2016, rev. 1.0.
[32] HDC2010 Low Power Humidity and Temperature Digital Sensors, Texas Instruments, Jul. 2017.
[33] MSP432P401R, MSP432P401M SimpleLink Mixed-Signal Microcontrollers, Texas Instruments, Mar. 2015, [Revised Mar 2017].
[34] GD5FxGQ4xBxIG DATASHEET., GigaDevice.
[35] J. Park, J. Dvok, and J. Havlk, “Device for long term measurement of heart rate,” Oct 2011.
[36] F. G. Yanowitz, “Introduction to ECG interpretation,” LDS Hospital and Intermountain Medical Center, 2012.
[37] C.-H. Lin, “Frequency-domain features for ECG beat discrimination using grey relational analysis-based classifier,” Computers Mathematics with Applications, vol. 55, pp. 680–690, Feb 2008.
[38] Analysis and interpretation of the electrocardiogram. Queen’s University. [Online]. Available: https://meds.queensu.ca/central/assets/modules/ECG/normal ecg.html [39] O. Inbar, A. Oren, M. Scheinowitz, A. Rotstein, R. Dlin, and R. Casaburi, “Normal cardiopulmonary responses during incremental exercise in 20- to 70-yr-old men,” Medicine and science in sports and exercise, vol. 26, pp. 538–46, Jun 1994.
[40] S. Alam and R. Gupta, “A DPCM based electrocardiogram coder with thresholding for real time telemonitoring applications,” Apr 2014.
[41] P. Zoran, D. Denic, J. Nikolic, A. Joci, and A. Jovanovic, “DPCM quantizer adaptation method for efficient ECG signal compression,” Journal of Communications Technology and Electronics, vol. 58, Dec 2013.
[42] J. Cox, F. M Nolle, H. A Fozzard, and G. C Oliver, “AZTEC, a preprocessing program for real time ECG rhythm analysis,” IEEE transactions on bio-medical engineering, vol. 15, pp. 128–9, May 1968.
[43] B. Furht and A. Perez, “An adaptive real-time ECG compression algorithm with variable threshold,” IEEE transactions on bio-medical engineering, vol. 35, pp. 489–94, Jul 1988.
[44] B. R. Shankara Reddy and I. S. N. Murthy, “ECG data compression using fourier descriptors,” IEEE transactions on bio-medical engineering, vol. BME-33, pp. 428 – 434, May 1986.
[45] Z. Lu, D. Youn Kim, and W. Pearlman, “Wavelet compression of ECG signals by the set partitioning in hierarchical trees algorithm,” IEEE transactions on bio-medical engineering, vol. 47, pp. 849 – 856, Aug 2000.
[46] L. Brechet, M.-F. Lucas, C. Doncarli, and D. Farina, “Compression of biomedical signals with mother wavelet optimization and best-basis wavelet packet selection,” IEEE transactions on bio-medical engineering, vol. 54, pp. 2186–92, Jan 2008.
[47] G.-A. Luo, S. L. Chen, and T.-L. Lin, “VLSI implementation of a lossless ECG encoder design with fuzzy decision and two-stage huffman coding for wireless body sensor network,” Dec 2013, pp. 1–4.
[48] E. Chua and W.-C. Fang, “Mixed bio-signal lossless data compressor for portable brainheart monitoring systems,” IEEE Transactions on Consumer Electronics, vol. 57, pp. 267 – 273, Mar 2011.
[49] D. Chacko John and Y. Lian, “A low complexity lossless compression scheme for wearable ECG sensors,” Jul 2015, pp. 449–453.
[50] T.-H. Tsai andW.-T. Kuo, “An efficient ECG lossless compression system for embedded platforms with telemedicine applications,” IEEE Access, vol. PP, pp. 1–1, Jul 2018.
[51] Q. Li, R. G Mark, and G. D Clifford, “Robust heart rate estimation from multiple asynchronous noisy sources using signal quality indices and a kalman filter,” Physiological measurement, vol. 29, pp. 15–32, Fab 2008.
[52] Z. Zhao and Y. Zhang, “SQI quality evaluation mechanism of single-lead ECG signal based on simple heuristic fusion and fuzzy comprehensive evaluation,” Frontiers in Physiology, vol. 9, Jun 2018.
[53] (2017) The schematic diagram of turn over. Golden Crown Orthopedic. [Online]. Available: https://kknews.cc/zh-tw/health/9m6xvvq.html