Google Cartographer

Google's open-source real-time lidar SLAM system, supporting both 2D and 3D mapping.

Cartographer is a real-time SLAM (simultaneous localization and mapping) system Google open-sourced in 2016, described in the accompanying paper by Hess et al., “Real-Time Loop Closure in 2D LIDAR SLAM.” It uses lidar (optionally combined with an IMU and odometry) to build local submaps, localizes through scan matching, and uses branch-and-bound search for loop-closure detection to eliminate the drift that accumulates over long runs, supporting both 2D and 3D. It connects to ROS through cartographer_ros, and was once one of the most common lidar mapping choices for cleaning robots, service robots, and ROS teaching. According to community feedback, the project has seen fewer updates in recent years, and ROS 2 users have often moved to SLAM Toolbox instead.

ExampleDrive a TurtleBot fitted with a 2D lidar around an office once, run Cartographer to produce an occupancy grid map, and hand that map to Nav2 for navigation.

Also called

cartographer_ros

Related

LiDAR SLAM · Simultaneous Localization and Mapping (SLAM) · Occupancy Grid Map · Loop Closure Detection · 2D LiDAR · Common ROS SLAM Packages (GMapping / SLAM Toolbox / RTAB-Map)

Sources

cartographer-project/cartographer (GitHub)
Cartographer 文档 (Chinese)