Walking characteristics extraction and behavior patterns estimation by using similarity with human motion map

Abstract

Recently development of autonomous control robot in human living space has been focused. It is necessary for the robot to decide its motion autonomously in this kind of space. Not only human position information but also human motion estimation is needed for the robot to decide its motion. In this study with observation experiment in various environments, human walking behavior is extracted. Observation experiment is made where can show human walking behavior obviously, and human position displacement is extracted as walking characteristic. An observation experiment was designed in different environment conditions where human walking characteristics could appear obviously, and human position displacement was calculated from the information of observed target. Human walking characteristics were extracted in each environment condition based on the calculated position displacement.