slam_toolbox

Slam Toolbox

We've received feedback from users and have robots operating in the following environments with SLAM Toolbox:

• Retail
• Warehouses
• Libraries
• Research

It is also the currently supported ROS2-SLAM library. See tutorials for working with it in ROS 2 Nav2 here.

Cite This Work

You can find this work here and clicking on the image below.

Macenski, S., Jambrecic I., "SLAM Toolbox: SLAM for the dynamic world", Journal of Open Source Software, 6(61), 2783, 2021.

Macenski, S., "On Use of SLAM Toolbox, A fresh(er) look at mapping and localization for the dynamic world", ROSCon 2019.

Introduction

Slam Toolbox is a set of tools and capabilities for 2D SLAM built by Steve Macenski while at Simbe Robotics, maintained while at Samsung Research, and largely in his free time.

This project contains the ability to do most everything any other available SLAM library, both free and paid, and more. This includes:

• Ordinary point-and-shoot 2D SLAM mobile robotics folks expect (start, map, save pgm file) with some nice built in utilities like saving maps
• Continuing to refine, remap, or continue mapping a saved (serialized) pose-graph at any time
• life-long mapping: load a saved pose-graph continue mapping in a space while also removing extraneous information from newly added scans
• an optimization-based localization mode built on the pose-graph. Optionally run localization mode without a prior map for "lidar odometry" mode with local loop closures
• synchronous and asynchronous modes of mapping
• kinematic map merging (with an elastic graph manipulation merging technique in the works)
• plugin-based optimization solvers with a new optimized Google Ceres based plugin
• RVIZ plugin for interacting with the tools
• graph manipulation tools in RVIZ to manipulate nodes and connections during mapping
• Map serialization and lossless data storage
• ... more but those are the highlights

This package has been benchmarked mapping building at 5x+ real-time up to about 30,000 sq. ft. and 3x real-time up to about 60,000 sq. ft. with the largest area (I'm aware of) used was a 200,000 sq. ft. building in synchronous mode (i.e. processing all scans, regardless of lag), and much larger spaces in asynchronous mode.

The video below was collected at Circuit Launch in Oakland, California. Thanks to Silicon Valley Robotics & Circuit Launch for being a testbed for some of this work.

An overview of how the map was generated is presented below:

1. ROS Node: SLAM toolbox is run in synchronous mode, which generates a ROS node. This node subscribes to laser scan and odometry topics, and publishes map to odom transform and a map.
2. Get odometry and LIDAR data: A callback for the laser topic will generate a pose (using odometry) and a laser scan tied at that node. These PosedScan objects form a queue, which are processed by the algorithm.
3. Process Data: The queue of PosedScan objects are used to construct a pose graph; odometry is refined using laser scan matching. This pose graph is used to compute robot pose, and find loop closures. If a loop closure is found, the pose graph is optimized, and pose estimates are updated. Pose estimates are used to compute and publish a map to odom transform for the robot.
4. Mapping: Laser scans associated with each pose in the pose graph are used to construct and publish a map.

Support and Contribution

If you have any questions on use or configuration, please post your questions on Robotics Stack Exchange with the slam and ros2 tags and someone from the community will work their hardest to get back to you. Tangible issues in the codebase or feature requests should be made with GitHub issues.

If you're interested in contributing to this project in a substantial way, please file a public GitHub issue on your new feature / patch. If for some reason the development of this feature is sensitive, please email the maintainers at their email addresses listed in the package.xml file.

For all contributions, please properly fill in the GitHub issue and PR templates with all necessary context. All PRs must be passing CI and maintaining ABI compatibility within released ROS distributions. A maintainer will follow up shortly thereafter.

03/23/2021 Note On Serialized Files

As of 03/23/2021, the contents of the serialized files has changed. For all new users after this date, this regard this section it does not impact you.

If you have previously existing serialized files (e.g. not pgm maps, but .posegraph serialized slam sessions), after this date, you may need to take some action to maintain current features. Unfortunately, an ABI breaking change was required to be made in order to fix a very large bug affecting any 360 or non-axially-mounted LIDAR system.

This Discourse post highlights the issues. The frame storing the scan data for the optimizer was incorrect leading to explosions or flipping of maps for 360 and non-axially-aligned robots when using conservative loss functions. This change permanently fixes this issue, however it changes the frame of reference that this data is stored and serialized in. If your system as a non-360 lidar and it is mounted with its frame aligned with the robot base frame, you're unlikely to notice a problem and can disregard this statement. For all others noticing issues, you have the following options:

• Use the <distro>-devel-unfixed branch rather than <distro>-devel, which contains the unfixed version of this distribution's release which will be maintained in parallel to the main branches to have an option to continue with your working solution
• Convert your serialized files into the new reference frame with an offline utility
• Take the raw data and rerun the SLAM sessions to get a new serialized file with the right content

More of the conversation can be seen on tickets #198 and #281. I apologize for the inconvenience, however this solves a very large bug that was impacting a large number of users. I've worked hard to make sure there's a viable path forward for everyone.

Multi-Robot SLAM

decentralized_multirobot_slam_toolbox_node extends slam_toolbox for multi-robot mapping. Each robot runs its own slam_toolbox instance under a unique namespace; localized scans are exchanged to align peer pose graphs over a shared global frame.

See docs/decentralized_multi_robot_slam.md for details on:

• Multi-Robot mapping
• Decentralized multi-robot slam architecture
• How to set up the shared global frame
• What topics are shared (and why)
• Example launch files & demo package

LifeLong Mapping

LifeLong mapping is the concept of being able to map a space, completely or partially, and over time, refine and update that map as you continue to interact with the space. Our approach implements this and also takes care to allow for the application of operating in the cloud, as well as mapping with many robots in a shared space (cloud distributed mapping). While Slam Toolbox can also just be used for a point-and-shoot mapping of a space and saving that map as a .pgm file as maps are traditionally stored in, it also allows you to save the pose-graph and metadata losslessly to reload later with the same or different robot and continue to map the space.

Our lifelong mapping consists of a few key steps

• Serialization and Deserialization to store and reload map information
• KD-Tree search matching to locate the robot in its position on reinitialization
• pose-graph optimization based SLAM with 2D scan matching abstraction

This will allow the user to create and update existing maps, then serialize the data for use in other mapping sessions, something sorely lacking from most SLAM implementations and nearly all planar SLAM implementations. Other good libraries that do this include RTab-Map and Cartographer, though they themselves have their own quirks that make them (in my opinion) unusable for production robotics applications. This library provides the mechanics to save not only the data, but the pose graph, and associated metadata to work with. This has been used to create maps by merging techniques (taking 2 or more serialized objects and creating 1 globally consistent one) as well as continuous mapping techniques (updating 1, same, serialized map object over time and refining it). The major benefit of this over RTab-Map or Cartographer is the maturity of the underlying (but heavily modified) open_karto library the project is based on. The scan matcher of Karto is well known as an extremely good matcher for 2D laser scans and modified versions of Karto can be found in companies across the world.

Slam Toolbox supports all the major modes:

• Starting from a predefined dock (assuming to be near start region)
• Starting at any particular node - select a node ID to start near
• Starting in any particular area - indicate current pose in the map frame to start at, like AMCL

In the RVIZ interface (see section below) you'll be able to re-localize in a map or continue mapping graphically or programmatically using ROS services.

On time of writing: there a highly experimental implementation of what I call "true lifelong" mapping that does support the method for removing nodes over time as well as adding nodes, this results in a true ability to map for life since the computation is bounded by removing extraneous or outdated information. Its recommended to run the non-full LifeLong mapping mode in the cloud for the increased computational burdens if you'd like to be continuously refining a map. However a real and desperately needed application of this is to have multi-session mapping to update just a section of the map or map half an area at a time to create a full (and then static) map for AMCL or Slam Toolbox localization mode, which this will handle in spades. The immediate plan is to create a mode within LifeLong mapping to decay old nodes to bound the computation and allow it to run on the edge by refining the experimental node. Continuing mapping (lifelong) should be used to build a complete map then switch to the pose-graph deformation localization mode until node decay is implemented, and you should not see any substantial performance impacts.

Localization

Localization mode consists of 3 things:

• Loads existing serialized map into the node
• Maintains a rolling buffer of recent scans in the pose-graph
• After expiring from the buffer scans are removed and the underlying map is not affected

Localization methods on image map files has been around for years and works relatively well. There has not been a great deal of work in academia to refine these algorithms to a degree that satisfies me. However SLAM is a rich and well benchmarked topic. The inspiration of this work was the concept of "Can we make localization, SLAM again?" such that we can take advantage of all the nice things about SLAM for localization, but remove the unbounded computational increase.

To enable, set mode: localization in the configuration file to allow for the Ceres plugin to set itself correctly to be able to quickly add and remove nodes and constraints from the pose graph, but isn't strictly required, but a performance optimization. The localization mode will automatically load your pose graph, take the first scan and match it against the local area to further refine your estimated position, and start localizing.

To minimize the amount of changes required for moving to this mode over AMCL, we also expose a subscriber to the `/ini