Development of vision based person following module for mobile robots in/out door environment

This paper proposes that a vision based robust person following method under varying illumination. The person following method is using a color stereo camera and a laser range sensor. The person detected method is based on disparity images and HSV color spaces from a color stereo camera and distance information from the laser range sensor. The developed person following method applies for an inverted pendulum type robot (Segway RMP) and a humanoid type service robot (enon). The validity of the proposed methods is confirmed though the person following experiments in in/out door environment.