A Model of Passive Human Motion Recognition Using Two-Layer Wireless Links

The recognition of human motions based on wireless sensor network technology has several obvious advantages. Since traditional recognition technology adopts a single wireless link, it is effective to identify whether an individual stays in a certain area. However, the technology fails to effectively recognize high-level human motions such as falling and lying down, which is of great importance for home health care for elders and rescue for fire buildings. To solve this problem, we put forward an experimental model with two layer wireless links to facilitate human motion recognition. After presenting relevant feasibility analysis, we conducted simulation experiments and applied Support Vector Machine (SVM) in the recognition of human motions. Experimental results show that the model makes a satisfactory performance in the recognition of the four human motions: walking, standing, lying down and normal state.