Artificial Marker-Aided Localization for Service Robots in Visually Repetitive Environments

In recent years, simultaneous localization and mapping (SLAM) has emerged as a critical technique for enabling the autonomous capabilities of service robots. However, practical SLAM applications often face significant performance challenges in visually and geometrically repetitive environments, which impede accurate localization. To address these challenges and ensure reliable localization for service robots, this work utilizes artificial markers, specifically AprilTags and ArUco tags. This paper presents a method that strategically leverages these markers, using them in minimal quantities while maximizing their localization impact. By leveraging the Kimera-VIO and TagSLAM algorithms, the proposed approach offers several contributions: (a) consistent pose estimation through a lightweight pose-graph optimizer suitable for resource-constrained systems; (b) an efficient strategy for optimizing the placement of artificial markers within target environments; and (c) validation of the proposed method in real-world environments, including warehouses and offices, using a commercially available service robot. This approach provides a viable industrial solution for improving localization performance in commercial service robots.