臺灣博碩士論文加值系統

行為辨識在智慧環境的相關研究是相當重要的一個議題。在實際的智慧環境中，一般會配置各種不同的感測器來偵測環境狀態與使用者的行動，智慧環境系統則根據感測器所偵測到的訊號來辨識使用者行為，進而提供合適的服務以滿足使用者的需求。因此，高正確率的行為辨識機制在智慧環境中扮演著相當重要的決定角色。然而，在智慧環境中，多重感測器所產生的資料是即時且連續的訊號，其中可能夾雜雜訊以及不規律的資料狀態，所以在多重感測器資料串流中，高正確率的行為辨識方法是相當困難的研究主題。在以往的研究中，隱藏式馬可夫模型 (Hidden Markov Model, HMM) 與條件隨機場 (Conditional Random Field, CRF) 是最常被使用來辨識序列資料的理論方法。本論文提出使用資料串流上序列樣式分析方法來建立行為辨識模型，以辨識智慧環境中使用者的行為。首先，我們會從已標記的資料串流中萃取行為序列樣式與行為轉換樣式並計算每一個序列樣式的分辨力。接著，本論文提出有效率的行為辨識方法來辨識資料串流上使用者的行為。驗證實驗則是使用WSU資料集與Kasteren資料集來測試本論文所提出的方法之效能並與隱藏式馬可夫模型做效能的比較。

Activity recognition is an important issue in the research area of smart environments. In real world, a smart environment is equipped with heterogeneous sensors to detect the status of the surroundings and the user’s actions. A smart environment can automatically provide proper services to satisfy the user’s requirements according to the user’s activities. Therefore, the high accurate activity recognition scheme is a critical process in the smart environment. However, the streaming data generated by sensors is real-time and on-line. The previous models including hidden Markov model (HMM) and conditional random field (CRF) are the principal theorems and methods to recognize sequential data. In this thesis, we propose a feature based activity recognition method for analyzing the streaming data. First, the activity patterns and transition patterns are extracted from the labeled streaming data and the discriminant coefficient of each activity pattern is computed. Then, the effective recognizing algorithms are developed for activity recognition. Finally, WSU dataset and Kasteren dataset were used to test our proposed methods and compared with the performance of the HMM.

論文摘要 I
ABSTRACT II
致謝 III
表目錄 V
圖目錄 VII
第1章 簡介 1
1.1 研究背景 1
1.2 研究動機與目的 2
1.3 論文貢獻 3
1.4 論文架構 3
第2章 序列分類的文獻探討 4
2.1 基於模型的序列分類 4
2.2 基於特徵的序列分類 6
第3章 行為辨識方法 9
3.1 行為辨識的系統架構 9
3.2 符號定義 10
3.3 行為序列樣式的萃取 11
3.4 計算分辨力係數 12
3.5 行為轉換的樣式萃取 17
3.6 行為辨識 18
第4章 實驗結果與討論 24
4.1 資料集介紹 24
4.2 評估方法的介紹 26
4.3 實驗一：離線學習 27
4.4 實驗二：漸進式學習 37
第5章 結論與未來研究方向 51
參考文獻 52
附錄 56

[1]http://www.icta.ufl.edu/gt.htm
[2]http://awarehome.imtc.gatech.edu/
[3]http://mavhome.uta.edu/
[4]http://www.cs.cmu.edu/~aura/
[5]http://www.mobihealth.org/
[6]http://www.medcan.com/
[7]http://www.eecs.harvard.edu/~mdw/proj/codeblue/
[8]http://www.ideatel.org/
[9]http://ailab.eecs.wsu.edu/casas/
[10]N. B. Bharatula, M. Stäger, P. Lukowicz, and G. Tröster, “Empirical Study of Design Choices in Multi-Sensor Context Recognition Systems,” in Proceedings of 2nd International Forum on Applied Wearable Computing, Zurich, Switzerland, 2005, pp. 79-93.
[11]L. Breiman, “Bagging Predictors,” Journal of Machine Learning, vol. 24, no. 2, 1996, pp. 123-140.
[12]N. A. Chuzhanova, A. J. Jones, and S. Margetts, “Feature Selection for Genetic Sequence Classification,” Journal of Bioinformatics, vol. 14, no.2, 1998, pp. 139-143.
[13]D. J. Cook, and M. Schmitter-Edgecombe, “Assessing the Quality of Activities in a Smart Environment,” Methods of Information in Medicine, vol. 48, no. 5, 2009, pp. 480-485.
[14]J. Fogarty, C. Au, and S. E. Hudson, “Sensing from the Basement: A Feasibility Study of Unobtrusive and Low-Cost Home Activity Recognition,” in Proceedings of the 19th Annual ACM Symposium on User Interface Software and Technology, Montreux, Switzerland, 2006, pp. 91-100.
[15]R. Hamid, S. Maddi, A. Bobick, and I. Essa, “Unsupervised Analysis of Activity Sequences Using Event-Motif,” in Proceedings of the 4th ACM International Workshop on Video Surveillance and Sensor Networks, Santa Barbara, CA, USA, 2006, pp.71-78.
[16]K. C. Hsu, Y. T. Chiang, G. T. Lin, and C. H. Lu, “Strategies for Inference Mechanism of Conditional Random Fields for Multiple-Resident Activity Recognition in a Smart Home” in Proceedings of the 23rd International Conference on Industrial, Engineering and Other Application of Applied Intelligent System, Cordoba, Spain, 2010, pp. 417-426.
[17]H. Junker, O. Amft, P. Lukowicz, and G. Tröster, “Gesture Spotting with Body-Worn Inertial Sensors to Detect User Activities,” Journal of Pattern Recognition, vol. 41, no. 6, 2008, pp. 2010-2024.
[18]T.L.M. van Kasteren, A. Noulas, G. Englebienne, and B.J.A. Kröse, “Accurate Activity Recognition in a Home Setting,” in Proceedings of the 10th International Conference on Ubiquitous Computing, Seoul, South Korea, 2008, pp. 1-9.
[19]T.L.M. van Kasteren and G. Englebienne, and B.J.A. Kröse, “Recognizing Activities in Multiple Context Using Transfer Learning,” in Proceedings of AAAI Symposium on AI in Eldercare, Washington, DC, USA, 2008.
[20]T.L.M. van Kasteren, G. Englebienne, and B.J.A. Kröse, “Activity Recognition Using Semi-Markov Models on Real World Smart Home Datasets,” Journal of Ambient Intelligent and Smart Environments, vol. 2, no. 3, 2010, pp. 311-325.
[21]T.L.M. van Kasteren, G. Englebienne, and B.J.A. Kröse, “Human Activity Recognition from Wireless Sensor Network Data: Benchmark and Software,” Activity Recognition in Pervasive Intelligent Environment, Atlantis Press, 2011, pp. 165-186.
[22]A. M. Khan, Y. K. Lee, S. Y. Lee, and T. S. Kim, “A Triaxial Accelerometer-Based Physical-Activity Recognition via Augmented-Signal Features and a Hierarchical Recognizer,” IEEE Transition on Information Technology in Biomedicine, vol. 14, no. 5, 2010, pp. 1162-1172.
[23]D. Kudenko, and H. Hirsh, “Feature Generation for Sequence Categorization,” in Proceedings of the 15th National Conference on Artificial Intelligence, Madison Wisconsin, USA, 1998, pp. 733-738.
[24]Y. X. Lin, and B. C. Chein, “A discriminant Based Document Analysis for Text Classification,” in Proceedings of International Computer Symposium, Workshop of Artificial Intelligence, Knowledge Discovery, and Fuzzy System, Tainan, Taiwan, 2010, pp. 594-599.
[25]R. Meir, and G. Rätsch, “An Introduction to Boosting and Leveraging,” Advanced Lectures on Machine Learning, vol. 2006, 2003, pp. 118-183.
[26]N. Lesh, M. J. Zaki, and M. Ogihara, “Mining Features for Sequence Classification,” in Proceedings of the 5th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Diego, CA, USA, 1999, pp. 342-346.
[27]J. Lester, T. Choudhury, N. Kern, G. Borriello, and B. Hannaford, “A Hybrid Discriminative/Generative Approach for Modeling Human Activities,” in Proceeding of the International Joint Conference on Artificial Intelligence, Edinburgh, Scotland, UK, 2005, pp. 766–772.
[28]N. Littlestone, “Learning Quickly When Irrelevant Attributes Abound: A new Linear-Threshold Algorithm,” Journal of Machine Learning, vol. 2, no. 8, 1988, pp. 285-318.
[29]L. R. Rabiner, “A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition,” in Proceedings of the IEEE, vol. 77, no 2, 1989, pp. 257-286.
[30]N. Ravi, N. Dandekar, P. Mysore, and M. L. Littman, “Activity Recognition from Accelerometer Data,” in Proceedings of the 20th International National Conference on Artificial Intelligence, Pittsburgh, Pennsylvania, CA, 2005, pp. 1541-1546.
[31]R. E. Schapire, and Y. Singer, “BoosTexter: A Boosting-Based System for Text Categorization,” Journal of Machine Learning, vol. 39, no. 2-3, 2000, pp. 135-168.
[32]E. M. Tapia, S. S. Intille, and K. Larson, “Activity Recognition in the Home Setting Using Simple and Ubiquitous Sensors” in Proceeding of the International Conference on Pervasive Computing, Linz, Austria, 2004, pp. 158–175.
[33]D. L. Vail, M. M. Veloso, and J. D. Lafferty, “Conditional Random Fields for Activity Recognition,” in Proceedings of the 6th International Joint Conference on Autonomous Agents and Multiagent Systems, Honolulu, Hawaii, 2007, pp. 1331-1338.
[34]J. A. Ward, P. Lukowicz, and G. Tröster, “Gesture Spotting Using Wrist Worn Microphone and 3-Axis Accelerometer,” in Proceedings of the 2005 Joint Conference on Smart Objects and Ambient Intelligence: Innovative Context-aware Service: Usages and Technologies, Grenoble, France, 2005, pp. 99-104.
[35]O. Yakhnenko, A. Silvesen, and V. Honavar, “Discriminatively Trained Markov Model for Sequence Classification,” in Proceedings of the International Conference on Data Mining, Houston, Texas, USA, 2005, pp. 498-505.