Detection and following of human legs using the SVDD (Support Vector Data Description) scheme for a mobile robot with a single Laser Range Finder

Abstract

The human-following is one of the significant procedure in human-friendly navigation of mobile robots. There are many potential applications of human-following technology. In this paper, we propose detection and tracking approaches for human legs by the use of a single LRF. We extract four simple attributes of human legs. To define the boundary of extracted attributes mathematically, we used a Support Vector Data Description (SVDD) scheme. When the target person moved with the close neighborhood of the other pedestrian, target tracker missed the target person. For this situation, we newly analyze the leg movement. The proposed approaches were successfully verified through various experiments.