Research on Localization and Relocalization of Unmanned Vehicles Based on Multi-Line LiDAR

With the rise of autonomous driving technology, the ability to accurately localize and re-localize unmanned vehicles in complex outdoor environments has become a critical challenge. This paper proposes a solution to this problem by integrating the Normal Distributions Transform (NDT) matching algorithm for localization and the Iterative Closest Point (ICP) algorithm for re-localization. The proposed methods are tested and validated in real-world outdoor settings, with experimental results showing that the NDT algorithm ensures localization accuracy of at least 0.3 meters, providing stable and reliable results. Furthermore, the ICP algorithm demonstrates high precision in vehicle re-localization, confirming its practical applicability and accuracy in dynamic environments.