Human-Leg Detection in 3D Feature Space for a Person-Following Mobile Robot Using 2D LiDARs

Abstract

People detection is an essential technique for person-following mobile robots in applications of human-friendly services and collaborative tasks. 2D light detection and ranging (LiDAR) sensors are useful for these applications, especially for applications that detect and follow people while maintaining a suitable distance, at a close range, using accurate range measurements. In this study, we propose a method of human-leg detection in 3D feature space for a person-following mobile robot equipped with a 2D LiDAR sensor. We also propose an improved LiDAR scan segmentation technique to extract segments of human leg candidates. The newly proposed method generates a feature vector with the attributes of leg shapes and learns a classification boundary in 3D feature space. Experimental results indicate that the proposed method successfully describes the target dataset and provides accurate leg detection. This study demonstrates that human legs can be detected with improved accuracy by learning the classification boundary in 3D feature space.