Cartographer

Google's real-time 2D and 3D SLAM system for multi-platform robotics

License

Apache-2.0

Overview

Cartographer is Google's open-source SLAM library that provides real-time simultaneous localization and mapping in 2D and 3D across multiple platforms and sensor configurations. It supports LiDAR, IMU, and odometry inputs, enabling robots to navigate and understand their surroundings without prior maps. Originally developed at Google, it became the most widely deployed open-source SLAM system in the ROS ecosystem.

Cartographer is Google's open-source SLAM library that builds real-time maps while simultaneously tracking a robot's position in 2D and 3D environments. It uses graph-based optimization with local and global SLAM components, supporting multiple sensor configurations including LiDAR, IMUs, and odometry for robust autonomous navigation.

Why it matters

Cartographer established the gold standard for open-source SLAM, becoming one of the most widely adopted mapping solutions in the ROS ecosystem. Its graph-based optimization approach and multi-trajectory mapping support made professional-grade SLAM accessible to the entire robotics community, from warehouse robots to autonomous vehicles.

Problem it solves

Robots need to simultaneously build a map of their environment and localize themselves within it. Cartographer provides a production-quality solution that handles sensor noise, loop closures, and multi-floor environments across various hardware platforms.

Key takeaways

• Production-grade 2D and 3D SLAM with loop closure detection
• Supports multiple sensor configurations (LiDAR, IMU, odometry)
• Graph-based pose optimization with local and global SLAM components
• Widely deployed reference implementation used by thousands of robots

Who it is for

• Robotics engineers building autonomous mobile robots
• Navigation stack developers working with ROS/ROS2
• Warehouse and logistics robotics teams
• Researchers studying SLAM and mapping algorithms

Limitations

• No longer actively maintained by the original Google team
• ROS2 fork (ros2/cartographer) recommended for new projects
• Configuration can be complex with many tuning parameters
• Real-time 3D SLAM requires significant computational resources