Human Tracking of Single Laser Range Finder Using Features Extracted by Deep Learning

Human recognition using single laser range finder (LRF) is utilized for the task of following a target person such as a cargo transport robot. In these recognition methods, the approach is applied in which human-crafted features is inputted to the one-class classification model to identify whether it is a human or not. In this paper, we propose a method that introduce features extracted by deep learning. In this method, we create an encoder that can extract features from input data using PointNet-based autoencoder. In its experiment, the features extracted by encoder is compared with the human-crafted features, and these extraction process length of time is measured.