Autonomous Mobile Mapping Robots: Key Software Components

Abstract

This chapter discusses key software components of autonomous mobile mapping robots equipped with an inertial measurement unit (IMU) and light detection and ranging (LiDAR). In recent years, new LiDARs with nonrepetitive scanning pattern have appeared in the market. They are also equipped with an IMU; thus, the front end of simultaneous localization and mapping (SLAM)—a robust LiDAR-inertial odometry framework—significantly improves unmanned ground vehicles (UGVs) and unmanned aerial vehicles (UAV) in 3D mapping scenarios. Our study incorporates FAST-LIO as the front end of SLAM. The main focus is a lightweight back-end implementation of pose graph simultaneous localization and mapping (SLAM). It is an alternative solution to state-of-the-art g2o or GTSAM implementations. We also elaborate on iterative closest point, normal distributions transform, and their extension for multiview 3D data registration/refinement. It is based on C++ using Eigen library. This chapter also discusses path planning in already mapped environment. All software components are available as open-source projects.