3D LiDAR-Based Localization Methods: An Overview

The localization problem is the basic problem of unmanned driving technology. In urban and indoor environments where satellite signals are unreliable, odometers become the key to the location system. The mechanical odometer cannot cope with the complexity of road conditions. Visual-based localization method is greatly affected by ambient light. More and more attention has been paid to LiDAR sensors because of their rich and dense spatial data and robustness to the environment. In this paper, the LiDAR-based localization methods are classified and reviewed, and the accuracy and runtime are compared using KITTI Odometry dataset. Finally, the advantages and drawbacks of various method are summarized, and the future directions of research are outlined. Also, the public datasets commonly used in localization research are listed.