Research on Classification of Human Daily Activities Based on a Single Tri-Axial Accelerometer

In this paper, a multi-layers method with multi-parameters based on the characteristics of the human movements acceleration signals is proposed to recognize the human daily activities. We calculate some features of the acceleration signals that are less dependent on the individuals. The features are successfully used to divide signals into different groups which are related to the human daily activities. Our experiments demonstrate the effectiveness of the proposed method, and the overall recognition accuracy is higher than 90%.