Contents



Preface

Autonomous mobile mapping robots produce digitized maps of their environment and are typically equipped with lidar, a camera, an inertial measurement unit, odometry, and sometimes GNSS for precise positioning. One of the challenges for these machines is simultaneous localization and mapping (SLAM), which is regarded as a "chicken and egg" problem: on one hand, the robot needs to be aware of its trajectory in order to construct a map; on the other hand, it needs to know the map in order to perform localization. There are several types of robots: unmanned ground vehicles (UGV), unmanned aerial vehicles (UAV), and unmanned surface vehicles (USV). Recent research has tended to equip all these robots with the same sensory setups, enabling common SLAM

technology to be applied in all domains: air, ground, underground, and surface.

Air and surface applications can use GNSS and GPS efficiently, as there are usually no obstacles. The biggest challenge for autonomous mobile mapping ground robots is to combine all functionalities in a common framework, enabling missions to be executed without collisions with obstacles. Another important challenge is the limited mobility of wheeled robots. Recent work on ledge-climbing robots shows that these robots are capable of performing complex autonomous mapping missions, such as underground mining mapping, nuclear plant mapping, and many others. Thus, ledge-climbing robots are desirable for ground and underground autonomous mobile mapping tasks. In the air, the problem of obstacles affecting the mission is not present. Recent commercial drones are capable of collecting data autonomously or even performing SLAM on board. Challenges remain, such as underground mapping using drones. Fortunately, recent advances in on-board on-line lidar odometry provide efficient solutions for localization in hazardous environments.

This book consists of eight chapters in four sections. Following an introduction, the first section investigates alternative approaches to mobile 3D scanning. The second section considers key software components used for building autonomous mobile robots, including lidar odometry, loop closure, pose graph SLAM, map refinement, path planning, and coverage algorithms. The third section discusses multi-robot mapping approaches and scalable algorithms used in SLAM. The fourth section describes important urban search and rescue applications of aerial 3D mapping and automotive

I would like to thank Prof. Andreas Nuechter and Prof. Piotr Skczypczyński for their

**Janusz Będkowski**

Warsaw, Poland

Polish Academy of Science,

Institute of Fundamental Technological Research IPPT,

SLAM in urban environments.

detailed contributions on real-world experiments.
