Recognition of Human Activities' Signals by Geometric Features

CHEN Lei,YANG Jie,SHEN Hong-bin,WANG Shuang-quan
DOI: https://doi.org/10.3321/j.issn:1006-2467.2008.02.014
2008-01-01
Abstract:Recognition of human activities from sensor signals is a hot research topic in wireless sensor networks.A novel hierarchy method was proposed to recognize numerous human activities,e.g.walking,going downstairs,going upstairs,and running.All the sensor signals were obtained from accelerometers on waist in the practical environment.This paper chooses geometric features and energy characteristics of the signals which are hardly depending on individuals to recognize the activities.The experiments demonstrate the effectiveness of the proposed method based on the geometric features of signals and more than 80% high recognition accuracy has been observed.
What problem does this paper attempt to address?