ML-Based Fast and Precise Target Docking of Autonomous Mobile Robots for Intelligent Transportation Systems Using 2-D LiDAR

Autonomous mobile robots (AMRs) are widely used in automated logistics transportation tasks, which is one of the fundamental parts of building intelligent logistics systems to improve efficiency in dynamic manufacturing and warehouse environments. To improve production efficiency, a target docking system is one of the important technologies for AMRs to quickly and accurately transport materials such as racks, carts, and pallets. In this paper, we propose a robust target docking algorithm based on 2D LiDAR data for battery-powered AMRs. The proposed method consists of four steps: detection, localization, path planning, and path tracking for fast and precise target docking using relative position and orientation measurements of the target. In addition, we propose a novel detection method based on machine learning to quickly detect various targets in a dynamic environment, which consists of three modules: first classification, secondary classification, and multiple matching-based 2D point cloud registration. The proposed method using an event-driven architecture overcomes problems such as poor docking performance, low efficiency, high-power consumption, and high response time. Unlike most existing docking methods that only consider static targets, our proposed method also solves the moving target docking problem in dynamic and unstructured environments. Real robot experiments have been performed to verify the target docking performance of the existing and proposed methods.