| As the commonly used robot in the logistics, AGV (Automated Guided Vehicle) system, which has the advantages of flexibility, reliability, low cost, and high security, has widely applied to tobacco, banking, and paper industry. In recent years, with the development of the machine vision and computer technology, the AGV navigation technology based on the machine vision has become an important research topic in the intelligent robot field. So, many researchers have devoted themselves to the technology and gain some achievements. For reliable navigation in any environment, AGV must know its pose (position and orientation). Therefore, estimating the position of the AGV is one of the fundamental problems of mobile robotics. So, many researchers at home and abroad have designed various locating systems based on vision. This paper presents a navigation method based on a navigating mark. The main contributions of the paper are summarized as follows:1. The development and status of AGV are summarized. The navigation method has also been presented and validated in the paper.2. The PCNN (Pulse Coupled Neural Network) image processing method has applied to the algorithm of recognizing the navigating mark.3. The paper presents a self-localization algorithm in the environment with navigating marks, whose position is known. The algorithm, which can estimate the pose of AGV, is based on range measurement of a single navigating mark from two or more arbitrary points with monocular vision.4. The self-localization algorithm in the case of information loss is studied and presented. |
