Time-aware Costmap for Smoother and Less Disruptive AMR Navigation With ROS 2

Abstract

Autonomous mobile robots (AMRs) are increasingly deployed in industrial settings, where they perform various tasks that replace human labor, thereby boosting operational efficiency. In such smart manufacturing environments where multiple types of robots coexist, effective management of dynamic obstacles during AMR navigation is crucial. However, the standard ROS 2 navigation stack lacks built-in mechanisms specifically to handle dynamic obstacles. Many existing studies on dynamic obstacle avoidance are limited to local path planning, focusing primarily on real-time recognition and evasion based on sensor data. In this paper, we propose a time-aware costmap framework that leverages information about areas frequently occupied by dynamic obstacles, integrating it into the global costmap to enable smoother and less-disruptive navigation. The framework not only enhances AMR stability but also increases productivity in industrial environments. The time-aware costmap framework offers three key advantages: First, it enables the global planner to select routes that are smooth and minimally disruptive. Second, it integrates a custom layer featuring time-bound dynamic obstacles into the global costmap as a plugin, enabling seamless adaptation within existing navigation frameworks. Finally, it works effectively using only a 2D LiDAR sensor, minimizing hardware and software overhead. We validate the performance of the proposed framework in a Gazebo simulation environment modeled after a milk-run distribution system. The results demonstrate that the framework substantially improves safety and stability by lowering curvature, jerk, and collision risks. While this comes with a trade-off of about 20% lower throughput due to smoother detours, the framework’s emphasis on smoothness and safety offers greater practical value for reliable navigation in industrial environments.