基於無監督學習的居家老人行為識別及異常檢測研究

Behavior recognition and irregularity detection using unsupervised learning based on sensor data for home elderly

尚翠娟(Cuijuan Shang)

指導教授 ： 張志勇

共同指導教授 ： 石貴平(Kuei-Ping Shih)

淡江大學/工學院/資訊工程學系博士班/博士(2021年)

https://doi.org/10.6846/TKU.2021.00193

摘要

人口老齡化給整個社會帶來很多問題，如醫療資源及護理人員的嚴重短缺、護理費用高昂等。考慮到多數老人更喜歡在家中安享晚年，同時結合老齡化產業現現狀，居家養老成為大多數政府主推的養老模式。而居家照護也成為了一個熱門研究議題，受到了廣泛的關注。
行為識別在支持老人居家養老中起著重要的作用。現有研究提出的行為識別演算法，多是通過從感測器資料中提取特徵或模式來識別老年人的行為。然而，這些方法大多基於帶有標記的感測器資料，採用概率模型或監督學習的方法來識別行為。本研究針對智慧家居中的獨居老人，基於無標記的感測器資料，提出了一種基於無監督學習的行為識別算法(BIA)。本研究基於對老人行為的觀察，提出了三個老人行為的特徵，即事件順序相似性、時間長度相似性和時間間隔相似性。基於這些行為觀察的特徵，定義了兩種行為屬性，即事件位移和長條圖形狀相似度。根據這些特性，提出了無監督學習的行為識別算法（BIA）。最後，實驗結果表明，該方法在行為識別精度和召回率方面優於現有的無監督機器學習機制。
老年人日常行為的異常檢測也是居家照護中的一個重要議題。現有的異常檢測研究多是基於某些生物醫學參數或某些特定行為的明顯異常來評估老年人的身體健康狀態。然而，很少有研究討論不同行為組合的隱性異常，這種隱性異常可以用來評估老年人的認知和身體健康，但不能基於感測器資料直接識別。因此，本研究提出一種隱式不規則檢測(IIRD)機制，旨在基於日常行為應用無監督學習算法來檢測老人行為為的隱式不規則性。本研究提出的IIRD機制能夠識別日常行為之間的距離和相似性，這是區分老年人日常行為的規律性和不規則性、檢測老年人健康狀況隱性不規則性的重要特徵。實驗結果表明，該方法在檢測準確率和Recall方面優於現有的無監督機器學習機制。
由於IIRD只輸出二值檢測結果，因此本研究進一步提出了一種基於特徵的隱式不規則性檢測機制(FIID)，該機制利用無監督學習提取規則性特徵，輸出隱式不規則性發生的概率。該方法將滿足時間規律性性和頻繁發生規律性性的規則行為作為識別為日常行為的規則特徵。這些特徵可以構造一個多維特徵空間來計算日常健康狀況的隱式不規則概率。實驗結果表明，本研究所提出的FIID在精度、Recall和F-measure方面都優於現有的隱式不規則機制。

關鍵字

行為識別 ； 異常檢測 ； 無監督學習 ； 傳感器數據 ； 居家照護

並列摘要

Advances in wireless sensor networks and increasing Internet-of-Things devices give great opportunities for smart homecare of the elderly. Smart homecare has been a promising issue and received much attention recently.
Behavior identification plays an important role in supporting homecare for the elderly living alone. In literature, plenty of algorithms have been designed to identify behaviors of the elderly by learning features or extracting patterns from sensor data. However, most of them adopted probabilistic models or supervised learning to identify behaviors based on labeled sensor data. This study proposes a behavior identification algorithm (BIA) using unsupervised learning based on unlabeled sensor data for the elderly living alone in smart home. This study presents the observation of elder behaviors with three features: Event Order, Time Length Similarity and Time Interval Similarity features. Based on these features of behavior observations, two properties of behaviors, including the Event Shift and Histogram Shape Similarity properties, are presented. According to these properties, the BIA is developed. Finally, performance results show that the BIA outperforms the existing unsupervised machine learning mechanisms in terms of the behavior identification precision and recall.
The irregularity detection of daily behaviors for the elderly also is an important issue in homecare. Plenty of mechanisms have been developed to detect the health condition of the elderly based on the explicit irregularity of several biomedical parameters or some specific behaviors. However, few research works focus on detecting the implicit irregularity involving the combination of diverse behaviors, which can assess the cognitive and physical wellbeing of elders but cannot be directly identified based on sensor data. This study proposes an Implicit IRregularity Detection (IIRD)mechanism that aims to detect the implicit irregularity by developing the unsupervised learning algorithm based on daily behaviors. The IIRD mechanism identifies the distance and similarity between daily behaviors, which are important features to distinguish the regular and irregular daily behaviors and detect the implicit irregularity of elderly health condition. Performance results show that the IIRD outperforms the existing unsupervised machine-learning mechanisms in terms of the detection accuracy and irregularity recall.
Since the IIRD simply outputs the binary detection results, this study proposes a feature-based implicit irregularity detection mechanism (FIID) which extracts the regularity features using unsupervised learning and outputs the probability of implicit irregularity. The FIID identifies the regular behaviors which satisfy the time-regular and happen-frequently properties as the regularity features of daily behaviors. These features then construct a multidimensional feature space to calculate the implicit irregularity probability of the daily health condition. Performance results show that the FIID outperforms the existing implicit irregularity mechanism in terms of precision, recall as well as F-measure.

並列關鍵字

Behavior recognition ； Irregularity detection ； Unsupervised learning ； Sensor data ； Homecare

參考文獻

[1] F. Zhou, J. Jiao, S. Chen and D. Zhang, "A case-driven ambient intelligence system for elderly in-home assistance applications," IEEE Trans. Syst. Man and Cybern., vol. 41, no. 2, pp. 179-189, Mar. 2011.

[2] J. Rafferty, C. D. Nugent, J. Liu and L. Chen, "From activity recognition to intention recognition for assisted living within smart homes," IEEE Trans. On Human-Machine Systems, vol. 47, no. 3, pp. 368-379, June 2017.

[3] C. Tsirmpas, A. Anastasiou, P. Bountris, and D. Koutsouris, "A new method for profile generation in an intemet of things environment: An application in ambient-assisted living," IEEE Internet of Things J., vol. 2, no. 6, pp. 471-478, Dec. 2015.

[4] H. Habibzadeh, et al., "A survey of healthcare Internet of Things (HIoT): a clinical perspective," IEEE Internet Things J., vol. 7, no. 1, pp.53-71, Jan. 2020.

[5] D. Naranjo-Hernández, L. M. Roa, J. Reina-Tosina and M. Á. Estudillo-Valderrama, "SoM: A smart sensor for human activity monitoring and assisted healthy ageing," IEEE Trans. Biomed. Eng., vol. 59, no. 11, pp. 3177-3184, Nov. 2012.