Limits...
Robust Indoor Human Activity Recognition Using Wireless Signals.

Wang Y, Jiang X, Cao R, Wang X - Sensors (Basel) (2015)

Bottom Line: Then, an online filtering method is designed to make actions' CSI curves smooth and allow them to contain enough pattern information.Each primitive action pattern can be segmented from the outliers of its multi-input multi-output (MIMO) signals by a proposed segmentation method.Lastly, in online activities recognition, by selecting proper features and Support Vector Machine (SVM) based multi-classification, activities constituted by primitive actions can be recognized insensitive to the locations, orientations, and speeds.

View Article: PubMed Central - PubMed

Affiliation: School of Software, Dalian University of Technology, Dalian 116620, China. dlutwangyi@dlut.edu.cn.

ABSTRACT
Wireless signals-based activity detection and recognition technology may be complementary to the existing vision-based methods, especially under the circumstance of occlusions, viewpoint change, complex background, lighting condition change, and so on. This paper explores the properties of the channel state information (CSI) of Wi-Fi signals, and presents a robust indoor daily human activity recognition framework with only one pair of transmission points (TP) and access points (AP). First of all, some indoor human actions are selected as primitive actions forming a training set. Then, an online filtering method is designed to make actions' CSI curves smooth and allow them to contain enough pattern information. Each primitive action pattern can be segmented from the outliers of its multi-input multi-output (MIMO) signals by a proposed segmentation method. Lastly, in online activities recognition, by selecting proper features and Support Vector Machine (SVM) based multi-classification, activities constituted by primitive actions can be recognized insensitive to the locations, orientations, and speeds.

Show MeSH

Related in: MedlinePlus

An example of online segmentation for two actions. (a) Two actions’ LOF subplots; (b) The segmentation result for two actions’ CSIs subplots by our method.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4541929&req=5

sensors-15-17195-f003: An example of online segmentation for two actions. (a) Two actions’ LOF subplots; (b) The segmentation result for two actions’ CSIs subplots by our method.

Mentions: CSIs are similar to speech signals, which can be classified into three states: the silence state (SS), the transitional state (TS), and the action state (AS). An action’s CSIs will normally go under the five states of SS–TS–AS–TS–SS. Due to the mechanism of human bodies, their actions normally have time intervals. So, we can use a time threshold to segment in-place actions roughly. Then we use the K-Means [1] to get the centers of actions from anomalous points of CSIs, e.g., in Figure 3, where there are two actions and by K-means we can get two clusters. Then by the proposed pattern segmentation method, we can get each action’s pattern data. Finally, in order to distinguish walking activities and in-place activities, we can adopt a cumulative moving variance of CSIs with a threshold.


Robust Indoor Human Activity Recognition Using Wireless Signals.

Wang Y, Jiang X, Cao R, Wang X - Sensors (Basel) (2015)

An example of online segmentation for two actions. (a) Two actions’ LOF subplots; (b) The segmentation result for two actions’ CSIs subplots by our method.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4541929&req=5

sensors-15-17195-f003: An example of online segmentation for two actions. (a) Two actions’ LOF subplots; (b) The segmentation result for two actions’ CSIs subplots by our method.
Mentions: CSIs are similar to speech signals, which can be classified into three states: the silence state (SS), the transitional state (TS), and the action state (AS). An action’s CSIs will normally go under the five states of SS–TS–AS–TS–SS. Due to the mechanism of human bodies, their actions normally have time intervals. So, we can use a time threshold to segment in-place actions roughly. Then we use the K-Means [1] to get the centers of actions from anomalous points of CSIs, e.g., in Figure 3, where there are two actions and by K-means we can get two clusters. Then by the proposed pattern segmentation method, we can get each action’s pattern data. Finally, in order to distinguish walking activities and in-place activities, we can adopt a cumulative moving variance of CSIs with a threshold.

Bottom Line: Then, an online filtering method is designed to make actions' CSI curves smooth and allow them to contain enough pattern information.Each primitive action pattern can be segmented from the outliers of its multi-input multi-output (MIMO) signals by a proposed segmentation method.Lastly, in online activities recognition, by selecting proper features and Support Vector Machine (SVM) based multi-classification, activities constituted by primitive actions can be recognized insensitive to the locations, orientations, and speeds.

View Article: PubMed Central - PubMed

Affiliation: School of Software, Dalian University of Technology, Dalian 116620, China. dlutwangyi@dlut.edu.cn.

ABSTRACT
Wireless signals-based activity detection and recognition technology may be complementary to the existing vision-based methods, especially under the circumstance of occlusions, viewpoint change, complex background, lighting condition change, and so on. This paper explores the properties of the channel state information (CSI) of Wi-Fi signals, and presents a robust indoor daily human activity recognition framework with only one pair of transmission points (TP) and access points (AP). First of all, some indoor human actions are selected as primitive actions forming a training set. Then, an online filtering method is designed to make actions' CSI curves smooth and allow them to contain enough pattern information. Each primitive action pattern can be segmented from the outliers of its multi-input multi-output (MIMO) signals by a proposed segmentation method. Lastly, in online activities recognition, by selecting proper features and Support Vector Machine (SVM) based multi-classification, activities constituted by primitive actions can be recognized insensitive to the locations, orientations, and speeds.

Show MeSH
Related in: MedlinePlus