MGLT: Magnetic-Lead Global Localization and Tracking in Degenerated Repetitive Environments

摘要

Accurate global localization and tracking are essential ingredients for autonomous mobile robots (AMRs) operating in enclosed or partially enclosed repetitive environments (e.g., office corridors, industrial warehouses, and transportation centers). In such degenerated challenging environments, global navigation satellite system signals are unreliable or severely degraded. Existing infrastructure-based (e.g., beacons/QR codes-based) solutions suffer from inflexibility and high maintenance costs, while traditional geometric feature-based LiDAR/visual localization technologies may fail due to highly similar structures. The ambient magnetic field (MF) has shown its omnipresence and high distinctiveness at different location, which makes it gradually adopted for infrastructure-free AMR localization applications. Currently, existing MF-based localization methods are partly based on 3-D MF vector matching or MF sequential matching, which typically suffer from the orientation-dependency and similar-sequential-route limitations, respectively. To address the aforementioned challenges, this article proposes a novel probabilistic Magnetic-lead Global Localization and Tracking (MGLT) framework using the proposed rotation-invariant Magnetic-Horizontal, Vertical, and Total intensity (M-HVT) descriptor and 2-D laser scans for AMRs operating in repetitive environments. The proposed framework mainly consists of: 1) Two-step Initialization: Laser range and MF sequence-based matching; and 2) MF-based Pose Tracking: recursive multidimensional MF sequence-based matching. Extensive simulations and real-world experimental results validate that the proposed framework outperforms SOTA MF-based localization baselines in both global localization convergence and tracking accuracy.