臺灣博碩士論文加值系統

醫藥技術的進步和健康照護服務的發展，導致高齡人口逐年增加，伴隨著少子化趨勢，人口高齡化已成為一種全球性現象。老化現象導致高齡者生理機能退化，獨立生活的能力降低，生活需要特殊的健康照護和輔助。本研究期望將環境科技輔助生活(Ambient Assisted Living, AAL)導入居家環境，以輔助和支持高齡者生活獨立性，經由監控和監督服務提供每日生活上的輔助，以提升生活能力和生活品質。
收集高齡者日常生活動作的情境資訊是一項耗時且困難的工作，除了隱私問題外，不影響高齡者正常生活動作、持續且長期的監測以及取得足夠的資料量，都是收集困難的主要因素。為了解決上述困難和節省時間成本，本研究先運用模擬的方式快速產生大量且符合高齡者日常生活動作的情境資訊，並運用演算法進行驗證後，再進行真實資料的收集與驗證。
本研究運用普松分布(Possion Distribution)及狀態轉變機率做為模擬方法，以活動能力為輕度依賴的護理之家高齡者的各種日常生活動作統計資料作為基準，產生大量日常生活情境資訊，並使用支援向量機進行異常行為之訓練及資料庫建立。由研究結果可知，本研究提出之架構可成功進行動作辨識與異常行為的偵測，可用於輔助看護、照護提供者或家人掌握高齡者居家生活情形與日常行為，異常行為動作的偵測則可提供適當的照護和輔助，使居家老人安全、便利和健康的進行日常生活行為，以提升高齡者居家的生活品質。

The advancement in medicine and healthcare services have prolonged the life expectancy of the elderly and led the world population to be aging. Aging has become an important public health issue because of the degradation of physiological functions decreased the elder’s ability of independent living and this resulted in the elder’s activities of daily living (ADLs) require special supports and cares. This work aimed to introduce the ambient assisted living (AAL) into the home environment to improve the independency of elders. Through monitoring and supervising the ADLs, living supports and services can be provided to improve the elder’s independence of living and quality of life.
However, collecting the context information of ADLs is a time-consuming and difficult task. It is very challenging to perform continuing and long-term collection to obtain sufficient amount of data without influencing the normal ADLs of the elderly and inducing the privacy issue. In order to overcome the collecting difficulties, this study developed a technique to emulate the contexts of elder’s ADLs that can quickly simulate sufficient amount of data to provide a cost-effective solution. By implementing intelligent algorithm, the elder’s daily activities can be recognized and abnormal behavior can be detected.
In this work, the elder’s ADLs are simulated by using Poisson distribution and probability of state transition in which the ADLs contexts of 56 light-dependent elders are used as normal basis. Ten ADLs were selected and recognized by implementing Hidden Markov Models (HMMs). Support Vector Machine (SVM) was used to detect the abnormal behavior. According to the results, the techniques in activity recognition and abnormal behavior detection were successfully performed in the proposed architecture. The proposed healthcare infrastructure can be used to assist the caregivers and the family members to well understand the conditions of the residential elderly and provide adequate cares and supports to the elderly. The proposed methodology is applicable to be applied into home healthcare and monitoring system for the elderly. The normal behavior model and abnormalities detection mechanism provides various employment for the healthcare systems and enhances the quality of life for the elderly.

誌謝 i
中文摘要 ii
Abstract iv
目 錄 vi
圖 目 錄 viii
表 目 錄 x
第一章 前言 1
1.1 研究目標 4
1.2 主要貢獻 6
1.3 論文架構 7
第二章 文獻回顧 8
2.1 簡介 8
2.2 行為與動作的呈現 9
2.3 行為與動作的辨識 11
2.3.1 統計方法 12
2.3.1.1 隱藏式馬可夫模型 13
2.3.1.2 有限狀態機 15
2.3.1.3 貝氏網路 16
2.3.1.4 其他統計方法 17
2.3.2 電腦計算方法 18
2.3.2.1 類神經網路 18
2.3.2.2 資料探勘技術 19
2.3.2.3 支援向量機 20
2.3.2.4 模糊系統 20
2.4 行為動作異常偵測 21
2.4.1 統計方法 22
2.4.1.1 高斯混合模型 22
2.4.1.2 隱藏式馬可夫模型 22
2.4.1.3 其他統計模型 23
2.4.2 電腦計算方法 24
2.4.2.1 類神經網路 24
2.4.2.2 資料探勘技術 24
2.4.2.3 支援向量機 25
2.4.2.4 模糊系統 25
2.5 討論 26
2.5.1 行為動作辨識 26
2.5.2 行為動作異常偵測 29
第三章 異常偵測研究方法 32
3.1 高齡者日常生活動作情境資訊的模擬 36
3.2 動作辨識之研究方法 41
3.3 高齡者行為異常偵測之研究方法 43
3.3.1 支援向量機 44
3.3.2 居家老人行為異常偵測之研究方法 49
第四章 實驗設計與結果 55
4.1 日常生活動作情境資訊的模擬與動作辨識 56
4.1.1 動作資訊模擬與動作辨識之實驗設計 59
4.1.2 模擬過程與動作辨識之結果 62
4.2 高齡者行為異常偵測 71
4.2.1 運用支援向量機訓練之實驗設計 72
4.2.2 異常偵測結果 76
4.3 運用無線射頻辨識系統(RFID)驗證系統架構 81
4.3.1 建置無線射頻辨識系統之實驗設計 82
4.3.2 實驗結果 84
第五章 討論與結論 88
5.1 討論 89
5.1.2 動作辨識 92
5.1.3 異常偵測 94
5.1.4 無線射頻辨識系統實作實驗 96
5.2 結論 99
參考文獻 100

圖2.1 時序性行為模式示意圖 9
圖2.2 N=2的動作序列直方圖 10
圖2.3 二元感測器的時間序列呈現方式，以位置為例 10
圖2.4 隱藏式馬可夫模型示意圖 14
圖3.1 四大行為類型依活動頻率(Activity Frequency, AF)和活動時間(Activity Interval, AI)分類之行為分布圖 33
圖3.2 狀態轉變示意圖 37
圖3.3 狀態轉變示意圖 37
圖3.4 動作辨識步驟方塊圖 42
圖3.5支援向量機示意圖 43
圖3.6 資料點分群示意圖 44
圖3.7 資料點由輸入空間映射到特徵空間之示意圖 45
圖3.8 線性分隔超平面示意圖(其中圓圈之資料點為支援向量點) 46
圖3.9 不同屬性之異常資料點示意圖 50
圖3.10.1 單一動作(”清潔”)的正常模式示意圖 52
圖3.10.2 單一動作(”清潔”)的異常狀況示意圖 52
圖4.1 機率陣列示意圖 60
圖4.2.1 起始狀態轉變機率陣列 60
圖4.2.2 起始輸出機率陣列 61
圖4.3 模擬程式使用介面 62
圖4.4.1 平均執行頻率AF長條圖 64
圖4.4.2 平均執行時間AI長條圖 64
圖4.5.1 訓練後的狀態轉變機率陣列(45天) 65
圖4.5.2 訓練後的狀態轉變機率陣列(90天) 65
圖4.5.3 訓練後的狀態轉變機率陣列(180天) 66
圖4.5.4 訓練後的狀態轉變機率陣列(225天) 66
圖4.5.5 訓練後的狀態轉變機率陣列(365天) 66
圖4.6.1 訓練後的輸出機率陣列(45天) 67
圖4.6.2 訓練後的輸出機率陣列(90天) 67
圖4.6.3 訓練後的輸出機率陣列(180天) 67
圖4.6.4 訓練後的輸出機率陣列(225天) 68
圖4.6.5 訓練後的輸出機率陣列(365天) 68
圖4.7.1狀態轉變機率訓練前後的尤拉距離變化趨勢圖 69
圖4.7.2輸出機率訓練前後的尤拉距離變化趨勢圖 69
圖4.8 不同天數資料量動作辨識準確度趨勢圖 70
圖4.9.1 三維空間支援向量機分群圖(屬性:Occurrence v.s. AI，動作1:睡覺) 77
圖4.9.2 支援向量機訓練資料(屬性:Occurrence v.s. AI，動作1:睡覺) 77
圖4.10 三維空間不同核心支援向量機分群圖(屬性:Weekly Frequency v.s. AI，動作1:睡覺) 78
圖4.11 異常行為週報(第1週至第4週) 80
圖4.12 實驗環境 82
圖4.13 實驗人員位置變換圖(編號:2，性別:女) 84
圖4.14 位置資訊歷史紀錄 85
圖4.15.1 支援向量機線性分群圖(訓練前) 86
圖4.15.2支援向量機線性分群圖(訓練後) 86

表3.1 位置資訊與動作關聯表 41
表4.1 模擬資訊與文獻資料比較表 63
表4.2 不同天數資料量之動作辨識準確度 70
表4.3 動作執行之強制性權重因子 74
表4.4 動作辨識準確度陣列 85
表4.5 訓練前後的異常動作偵測準確度 87

[1] United Nations. "World Population Ageing 2013,” Department of Economic and Social Affairs, Population Division, pp.10, 2013.
[2] U.S. Census Bureau, International Population Reports WP/02," Global Population Profile: 2002", 2004.
[3] 內政部統計處,"簡易生命表函數定義及編算方法," 2014.
[4] 內政部統計處, "104 年國人零歲平均餘命估測結果," 2016.
[5] C. D. Mathers, R. Sadana, J. A. Salomon, C. J. Murray, A. D. Lopez. "Estimates of DALE for 191 countries: methods and results," Global Programme on Evidence for Health Policy Working Paper No. 16, 2000.
[6] 行政院主計總處. "國民幸福指數年報," 2016.
[7] 徐業良. "老人福祉科技與遠距居家照護技術," 滄海圖書股份有限公司出版，台中市，2008.
[8] R. Casas, R. Blasco Marn, A. Robinet, A. R. Delgado, A. R.Yarza, J. Mcginn, R. Picking and V. Grout. "User modelling in ambient intelligence for elderly and disabled people," Proceedings of the 11th International Conference on Computers Helping People with Special Needs, pp. 114-122, 2008.
[9] F. Rivera Illingworth, V. Callaghan, and H. Hagras. "Detection of normal and novel behaviours in ubiquitous domestic environments," The Computer Journal, Vol. 53, pp.142-151, 2010.
[10] K. Vrotsou, K. Ellegard, M. Cooper. "Everyday life discorveies:Mining and visualizing activity patterns in social science diary data," the 11th internation conference on Information Visualization 2007(IV'07), Switzerland, July 2007.
[11] Raffay Hamid, Siddhartha Maddi, Amos Johnson, Aaron Bo-bick, Irfan Essa, and Charles Isbell. "A novel sequence representation for unsupervised analysis of human activities," Elsevier Science, Artificial Intelligence, Vol. 173, No.14, pp.1221-1244, 2009.
[12] M. Javad Akhlaghinia, Ahmad Lotfi, Caroline Langensiepen, and Nasser Sherkat. "Occupant behaviour prediction in ambient intelligence computing environment," International Journal of Uncertain Systems, Vol. 2, No.2, pp.85-100, 2008.
[13] Sumi Helal, Eunju Kim and Diane Cook. "Human activity recognition and pattern discovery," IEEE Pervasive Computing, Vol. 9, pp.48-53, January 2010.
[14] Liming Chen and Ismail Khalil. "Activity recognition: Approaches, practices and trends," Activity Recognition in Pervasive Intelligent Environments, Vol. 4, pp.1-32, May 2011.
[15] S. M. Mahmoud. "Identification and Prediction of Abnormal Behaviour Activities of Daily Living in Intelligent Environments," Phd. Thesis, Nottingham Trent University, 2012.
[16] Juan Carlos Augusto. "Ambient intelligence: Basic concept and applications," International Conference on Software and Data Technologies, Communications in Computer and Information Science, pp.16-26, 2006.
[17] Leonard E. Baum and Ted Petrie, "Statistical Inference for Probabilistic Functions of Finite State Markov Chains," The Annals of Mathematical Statistics, Vol. 37, No. 6, 1966.
[18] John Lafferty, Andrew McCallum, Fernando Pereira, "Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data," ICML '01 Proceedings of the Eighteenth International Conference on Machine Learning, pp.282-289, 2001.
[19] Chiang, Chih-Yen; Liu, Kai-Chun; Hsu, Steen J.; Chan, Chia-Tai, "Compound Gesture Classification for Context-Aware Healthcare Monitoring System," Journal of Medical Imaging and Health Informatics, Vol. 4, No. 3, June 2014, pp. 427-432, 2014.
[20] Sean R. Eddy. "What is a hidden Markov model?" Nature Biotechnology, Vol. 22, pp.1315-1316, 2004.
[21] Martin Floeck, Lothar Litz, and Thorsten Rodner. "An ambient approach to emergency detection based on location tracking," In Proceedings of the 9th international conference on Toward useful services for elderly and people with disabilities: smart homes and health telematics, Springer-Verlag, ICOST'11, Berlin, Heidelberg, pp. 296-302, 2011.
[22] Y. Du, F. Chen, W. Xu, Y. Li, “Recognizing interaction activities using dynamic Bayesian network,” in: ICPR’06: Proceedings of the 18th International Conference on Pattern Recognition, IEEE Computer Society, Washington, DC, USA, 2006, pp. 618–621.
[23] L. Atallah, G-Z Yang. "The use of pervasive sensing for behaviour profilling-a survey," Pervasive Computing, Vol. 5, Issue 5, pp.447-464, 2009.
[24] Daniel Wilson and Chris Atkeson. "Automatic health monitoring using anonymous binary sensors," In Extended Proceedings of CHI 2004:Workshops Vienna, Austria, pp. 1719-1720, 2004.
[25] Tracy Barger, Donald Brown1, and Majd Alwan. "Health status monitoring through analysis of behavioural patterns," Springer-Verlag, In 8th Congress of the Italian Association for Articial Intelligence (AI*IA) on Ambient Intelligence, pp. 22-27, 2003.
[26] Kelvin Sim, Clifton Phua, Ghim-Eng Yap, Jit Biswas, and Mounir Mokhtari. "Activity recognition using correlated pattern mining for people with dementia," 33rd Annual International Conference of the IEEE EMBS Boston Massachusetts USA, 2011.
[27] Jack V. Tu, "Advantages and Disadvantages of Using Artificial Neural Networks versus Logistic Regression for Predicting Medical Outcomes," Journal of Clinical Epideminology, Vol. 49, No. 11, pp. 1225-1231, 1996.
[28] Diane J. Cook. "Making sense of sensor data," IEEE Pervasive Computing, Vol. 6, pp.105-108, 2007.
[29] Stefan Squartini, Arnir Hussain, and Francesco Piazza. "A recurrent multi-scale architecture for long-term memory prediction task," Speech,and Signal Processing, Vol. 2, Issue 6-10, pp.789-792, 2003.
[30] R. Illingworth, V. Callagha, and H. Hagras. "A neural network agent based approach to activity detection in ami environments," In IEE In-ternational Workshop, Intelligent Environments (IE05), pp. 1-12, 2005.
[31] Vikramaditya Jakkula. "Predictive data mining to learn health vitals of a resident in a smart home," Seventh IEEE International Conference on Data Mining Workshops, pp. 163-168, October 2007.
[32] George Papamatthaiakis, George C. Polyzos, and George Xy-lomenos. "Monitoring and modelling simple everyday activities of the elderly at home," IEEE Press, In Proceedings of the 7th IEEE conference on Consumer com-munications and networking conference, CCNC'10, pp. 617-621 ,2010.
[33] Corinna Cortes and Vladimir Vapnik. "Support-Vector Networks," Machine Learning, Vol. 20, Issue 3, pp. 273-297, 1995.
[34] Steve R. Gunn, 1998, "Support Vector Machine for Classification and Regression," Faculty of Engineering, Science and Mathematics School of Electronics and Computer Sciemce, 1998.
[35] Rachid Kadouche, Hélène Pigot, Bessam Abdulrazak Sylvain Giroux. "User's behaviour classification model for smart houses occupant prediction," Atlantis Ambient and Pervasive Intelligence, Vol. 4, pp.149-164, 2011.
[36] Hamid Medjahed, Dan Istrate, Jerome Boudy, and Bernadette Dorizzi. Human activities of daily living recognition using fuzzy logic for elderly home monitoring,” FUZZ IEEE Korea, pp. 20-24, August 2009.
[37] Faiyaz Doctor, Hani Hagras, and Victor Callaghan. "A fuzzy embedded agent-based approach for realizing ambient intelligence in intelligent inhabited environments," IEEE Transaction on Systems, Man, and Cybernetics-part A: Systems and Humans, Vol. 35, pp. 55-65, January 2005.
[38] Chandola, V., Banerjee, A., Kumar, V.. "Anomaly Detection: A Survey," ACM Computing Surveys, Vol. 14, Issue. 3, pp.1-72, 2009.
[39] Mark Hawley Fabien Cardinaux, Simon Brownsell and David Bradley. "Modelling of behavioural patterns for abnormality detection in the context of lifestyle reassurance," Proc. CIARP 2008, Havana, Cuba,pp. 243-251, 2008.
[40] Taketoshi Mori, Akinori Fujii, Masamichi Shimosaka, Hiroshi Noguchi, and Tomomasa Sato. "Typical behaviour patterns extraction and anomaly detection algorithm based on accumulated home sensor data," Future Generation Communication and Networking, Vol. 2, pp. 12-18, 2007.
[41] Wonjoon Kang, Dongkyoo Shin, Dongil Shin. "Detecting and predicting of abnormal behavior using hierarchical Markov model in smart home network," 2010 IEEE 17Th International Conference on Industrial Engineering and Engineering Management, 2010.
[42] Sebastian Luhr, Svetha Venkatesh, Geoff West, Hung H. Bui. "Duration Abnormality Detection in Sequences of Human Activity," Technical Report, Dept. of Computing, Curtin University of Technology, 2004.
[43] Diane J. Cook Gaurav Jain and Vikramaditya Jakkula. "Monitoring health by detecting drifts and outliers for a smart environment inhabitant," IOS Press, In 4th International Conference on Smart Homes and Health Telematics, pp. 114-121, June 2006.
[44] Colin O'Reilly, Alexander Gluhak, Muhammad Ali Imran. "Distributed Anomaly Detection using Minimum Volume Elliptical Principal Component Analysis," IEEE Transactions on Knowledge and Data Engineering, Vol. 28, Issue 9, pp. 2320-2333, 2016.
[45] George Stefatos, A. Ben Hamza. "Cluster pca for outliers detection in high-dimensional data," 2007 ISIC IEEE International Conference on Systems, Man and Cybernetics, Canada, 2007.
[46] F. R. Illingworth, V. Callagha, and H. Hagras. "Towards the detection of temporal behavioural patterns in intelligent environments," In 2nd IET International Conference on Intelligent Environments, pp. 119-125, 2006.
[47] Hsu Hui-Huang, Lu Kun-Chi, and M. Takizawa. "Abnormal behaviour detection with fuzzy clustering for elderly care," Computer Symposium, pp, 6-11, December 2010.
[48] J. Loane, B. O'Mullane, and B. Bortz andR.B. Knapp. "Interpreting presence sensor data and looking for similarities between homes using cluster analysis," 5th International Conference on Pervasive Computing Technologies for Healthcare (PervasiveHealth), pp. 438, May 2011.
[49] Vikramaditya Jakkula, Diane J. Cook, and Aaron S. Crandall."Temporal pattern discovery for anomaly detection in a smart home," In 3rd IET International Conference on Intelligent Environments, pp. 339-345, 2007.
[50] Jie Yin, Qiang Yang, and Jeffrey Junfeng Pan. "Sensor-based abnormal human-activity detection," IEEE Transaction on Knowledge and Data Engineering, Vol. 20, pp.1082-1090, August 2008.
[51] Jae Hyuk Shin, Boreom Lee, and Kwang Suk Park. "Detection of abnormal living patterns for elderly living alone using support vector data description," IEEE Transaction on Information Technology in Biomedicine, Vol. 15, pp.438-448, May 2011.
[52] Trevor Martin, Basim Majeed, Beum-Seuk Lee, and Nick Clarke. "Fuzzy ambient intelligence for next generation tele-care," IEEE In-ternational Conference on Fuzzy Systems; Canada, pp. 16-21, July 2006.
[53] Murray N. Rothbard. "The Logic of Action I: Method, Money, and the Austrian School," Edward Elgar Pub., pp. 58-77, 1997.
[54] Xiao-Pu Han, Tao Zhou, Bing-Hong Wang. "Modeling Human Dynamics with Adaptive Interest," New Journal of Physics, Vol. 10, July 2008.
[55] Jiawei Han, Micheline Kamber, Jian Pei. "Data Mining: Concepts and Techniques," 3rd edition, Morgan Kaufmann Publisher of Elsevier, USA, 2012.
[56] Michael Bridges, Elizabeth A. heron, Colm O'Dushlaine, Ricardo Segurado, Derek Morris, Aiden Corvin, Michael Gill, Carlos Pinto. "Genetic Classification of Populations Using Supervised Learning," The International Schizophrenia Consortium (ISC), Vol. 6, Issue 5, pp. 1-12, 2011.
[57] Zoubin Ghahramani. "Unsupervised Learning," the Lecture Notes in Computer Science book series (LNCS), Vol. 3176, pp. 72-112, 2004.
[58] Ling-Fang Wang, Szu-Yu Tzeng. "A Study on Daily Behaviour and Territory Activity for the Elderly in Nursing Home-Southern Taiwan's Two Hospital-Based Nursing Homes for Example," Journal of architecture, No. 57, pp.25-53, 2006.
[59] Hyeran Byun and Seong-Whan Lee. "Applications of Support Vector Machines for Pattern Recognition: A Survey," the Lecture Notes in Computer Science book series (LNCS), Vol. 2388, pp. 213-236, 2002.
[60] CHRISTOPHER J.C. BURGES. "A Tutorial on Support Vector Machines for Pattern Recognition,", Data Mining and Knowledge Discovery, Vol. 2, Issue 2, pp. 121–167, 1998.
[61] John Platt. "Sequential Minimal Optimization: A Fast Algorithm for Training Support Vector Machines," Technical Report, Microsoft Research, pp. 1-21, 1998.