A Design of Learning based Human Positioning Algorithm using Images

Abstract

In this paper, a learning based human positioning algorithm using image was proposed to overcome the disadvantages of fingerprint map-based localization on hardware devices, and the human positioning accuracy was also verified through an experiment. The core parts of proposed algorithm are the learning phase of human positioning map (HPM) and human positioning phase. As a result of the experiment, the human recognition accuracy of proposed algorithm showed a high recognition accuracy of over 99.74% when a person moves between Om and 6m.