Performance Analysis of 3D NDT Scan Matching for Autonomous Vehicles Using INS/GNSS/3D LiDAR-SLAM Integration Scheme

Abstract

Because robustness and accuracy of localization are crucial for autonomous driving applications. Using the conventional integration scheme of Inertial Navigation System (INS) and Global Navigation Satellite System (GNSS), pose estimation error can drift and accumulate with time, especially in GNSS challenging environment and in unknown environment where an existing map has not been constructed. In this paper, in term of using multi-sensor fusion for improving the positioning accuracy, we proposed a localization method that is based on LiDAR-based 3D Normal Distribution Transform (NDT) scan matching with an INS/GNSS integration scheme. As the experimental results, our proposed method showed a statistical improvement over the state of the art INS/GNSS integration scheme.