近年來因為現代人對健康的觀念越來越重視，以及台灣漸漸邁向高齡化的社會，看護成本逐年提高，因此較符合低成本、適合長時間監測的非接觸式生理訊號監測(Non-contact vital signs monitoring)技術的需求與日漸增。
為了研究如何利用非接觸式方法去探測基礎生理訊號頻率，本篇論文使用了都卜勒雷達(Doppler radar)系統做為工具，透過訊號採集以及訊號處理、頻譜分析的方式，試圖偵測出被測者的心跳頻率(Heart rate)。我們測試了四種方法，分別為離散傅立葉變換法(Discrete fourier transform)、相位能量法(Phase energy method)、速度強度法(Velocity strength method)、與循環頻譜(Cyclic spectrum)作為偵測心跳訊號頻率的核心演算法的選擇，並對這四種演算法進行抗雜訊能力以及運算複雜度進行分析，尋找出最適合用來偵測週期性生理訊號的核心演算法。最後再加上訊號統計量以及演算法結果過濾的步驟，進一步提出了我們的人體心跳偵測演算法(Human heart rate detection algorithm)，能透過心跳運動存在與否，判斷是否有人位於雷達前，並且能輸出估計的心律，達到監控人體的目的。
本研究也開發了一套以MATLAB GUI為基礎的介面程式，並設計一些實驗，去驗證我們的演算法的可靠性，包括是否有心跳運動存在的準確率(Accuracy)和其心律是否吻合的匹配率(Matching rate)等數值進行統計。另外，此介面也額外提供了即時偵測的功能。我們透過真實訊號的實驗結果證明，本研究所提出的演算法可以在一定距離內，準確偵測到靜止狀態下的人的心跳頻率。

Due to the fact that people pay more attention on health nowadays, and the society of Taiwan gradually towards to an aging one, the cost of health care are increasing year by year. Under the circumstances, the low cost, suitable for long-time monitoring purpose, such as non-contact vital signs monitoring technologies, are urgently demanded.
In order to detect human’s vital signs in non-contact way, this thesis adopts a Doppler radar system as the detection tool coupled with several signal processing approaches to estimate the heart rate of human. Four core methods are seleted to capture the periodic frequency of human heart motion, including Discrete fourier transform (DFT) method, Phase energy method, Velocity strength method, and Cyclic spectrum method. According to the performance comparison in both effectiveness and efficiency, we find that the Cyclic spectrum-based method perform the best. Therefore, we select Cyclic spectrum method as the main method, together with additional pre-processing and post-filtering steps, to propose a human’s heart rate detection algorithm, which can accurately determine if a human’s heart motion exist or not.
This study further develops a MATLAB GUI interface, and designs some scenario experiments to verify the reliability of our algorithm. For the performance evaluation，the GUI can output the accuracies and matching rates of the detection results once the groundtruth and reference information are available. Through the experiments conducted on real data collected by the Doppler radar, it can be proved that the proposed algorithm can reliably detect human’s heart rate when he or she stays in a stationary condition.

審定書 i
誌謝 ii
摘要 iii
Abstract iv
目錄 vi
圖目錄 viii
表目錄 xi
第一章 緒論 1
1.1 研究目的 1
1.2 文獻回顧 2
1.3 論文架構 3
第二章 都卜勒雷達簡介 5
2.1 都卜勒雷達基本原理 5
2.2 都卜勒雷達基本構造 7
第三章 心跳頻率偵測演算法 11
3.1 訊號處理流程圖 11
3.2 訊號前處理(Step.1) 13
3.3 核心演算法(Step.2) 15
3.3.1 離散傅立葉變換法 (Discrete Fourier Transform，DFT) 17
3.3.2 相位能量法(Phase energy method) 20
3.3.3 速度強度法(Velocity strength method) 24
3.3.4 循環頻譜法(Cyclic spectrum method) 27
3.4 演算法抗雜訊能力比較 35
3.5 篩選與驗證方法 47
3.5.1 篩選方法 47
3.5.2 驗證方法 51
第四章 實驗結果與統計 56
4.1 圖形使用者介面(GUI)開發 56
4.2 Scenario 1：人坐在電腦桌前 57
4.2.1 Scenario 1-1 57
4.2.2 Scenario 1-2 62
4.3 Scenario 2：人躺在沙發上睡覺 68
4.3.1 Scenario 2-1 68
4.3.2 Scenario 2-2 73
4.4 實驗結果統計圖表 79
4.5 實驗觀察與總結 82
第五章 結論與未來展望 85
5.1 結論 85
5.2 未來展望 86
參考文獻 87

[1] Gabriel Reyes, Di Wang, Rakesh Nair1, Changzhi Li, Xiaolin Li, and Jenshan Lin," A Doppler Radar Sensor Platform for Monitoring Activity Levels"
[2] Adrian Lin and Hao Ling," Through-Wall Measurements of a Doppler and Direction-of-Arrival (DDOA) Radar for Tracking Indoor Movers"
[3] Ehsan Yavari, Hsun Jou, Victor Lubecke, and Olga Boric-Lubecke," Doppler Radar Sensor for Occupancy Monitoring"
[4] DAI Shun, ZHU Fang, XU Yan-yun, FANG Guang-you," Vital Signal Detection Method Base on Principal Component Analysis and Empirical Mode Decomposition for Ultra Wideband Radar," ACTA ELECTRONICA SINICA Feb.2012 Vol.40 No.2.
[5] Nuno Miguel Santos Marques," Heart Sound Segmentation: A Stationary Wavelet Transform Based Approach," Faculdade de Ciências da Universidade do Porto, September 2013
[6] J. Antoni," Cyclic Spectral Analysis in Practice," Mechanical Systems and Signal Processing, Volume 21, Issue 2, February 2007, Pages 597-630.
[7] Somayeh Kazemi, Ayaz Ghorbani, Hamidreza Amindavar and Changzhi Li," Cyclostationary Approach for Heart and Respiration Rates Monitoring with Body Movement Cancellation Using Radar Doppler System ," Member, IEEE,8 Oct 2013.
[8] Mohammad Shaifur Rahman and Ki-Doo Kim," Extended Kalman Filter for Rate Estimation in Doppler Radar Cardiopulmonary Monitoring System," International Journal of Bio-Science and Bio-Technology Vol. 4, No. 4, December, 2012.
[9] Mohammad Shaifur Rahman and Ki-Doo Kim," Extended Kalman Filter for Rate Estimation in Doppler Radar Cardiopulmonary Monitoring System," International Journal of Bio-Science and Bio-Technology Vol. 4, No. 4, December, 2012.
[10] Eric April," The advantage of cyclic spectral analysis," Defence research establishment Ottawa technical note 92-4
[11] David M. Pozar," Microwave and RF design of wireless systems," John Wiley & Sons, Ltd., 2000.
[12] A.Napolitano," Cyclostationary Signal Processing and its Generalizations," June 29, 2014
[13] Chenyan Song, Ehsan Yavari1, Aditya Singh1, Olga Boric-Lubecke1 and Victor Lubecke," Detection Sensitivity and Power Consumption vs. Operation Modes Using System-on-Chip Based Doppler Radar Occupancy Sensor"
[14] Chen-Chia Wang, Sudhir B. Trivedi, Feng Jin, Serguei Stepanov, Zhongyang Chen, Jacob Khurgin, Ponciano Rodriguez, and Narasimha S. Prasad," Human Life Signs Detection Using High-Sensitivity Pulsed Laser Vibrometer," IEEE SENSORS JOURNAL, VOL. 7, NO. 9, SEPTEMBER 2007
[15] SangHyun Chang, Naoki Mitsumoto, Joel W. Burdick," An algorithm for UWB radar-based human detection"
[16] Sevgi Zübeyde Gürbüz, William L. Melvin, and Douglas B. Williams," Comparison of Radar-Based Human Detection Techniques"
[17] P. van Dorp and F.C.A. Groen," Human walking estimation with radar"
[18] Bijan G. Mobasseri and Moeness G. Amin," A Time-Frequency Classifier for Human Gait Recognition"
[19] Anders Høst-Madsen, Nicolas Petrochilos, Olga Boric-Lubecke, Victor M. Lubecke, Byung-Kwon Park, and Qin Zhou," Signal Processing Methods for Doppler Radar Heart Rate Monitoring"
[20] Amy C. Malady," Cyclostationarity Feature-Based Detection and Classification"
[21] Mohammad Shaifur Rahman and Ki-Doo Kim," Extended Kalman Filter for Rate Estimation in Doppler Radar Cardiopulmonary Monitoring System," International Journal of Bio-Science and Bio-Technology Vol. 4, No. 4, December, 2012