Skip to content

Commit

Permalink
Update README.md
Browse files Browse the repository at this point in the history
  • Loading branch information
fnobis committed Jul 17, 2020
1 parent 1bcd5aa commit 536e2d5
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion README.md
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
# ORB-SLAM2 map saving and loading for closed circuit localization

This project provides a ROS package for localization of an autonomous vehicle. The localization is based on a map consisting of ORB features. The mapping and localization module is taken from the <a href="https://github.com/raulmur/ORB_SLAM2">ORB-SLAM2</a> implementation. Our project builds on top of the [ROS-enabled version](https://gitlab.tu-berlin.de/breakdowncookie/ORB_SLAM2"). In this extensions the map of ORB-features be saved to the disk as a reference for future runs along the same track. At high speeds, a map-based localization, may be prefered over a full SLAM due to better localization accuracy and global positioning reference. The use-case for the presented system is a closed circuit racing scenario. In a first step, we create a map at low speeds with a full SLAM. In further runs at higher speeds, we localize on the previously saved map to increase the localization accuracy. The usage is demonstrated with an example from the KITTI dataset. The flow chart of the extended functionality is shown in Figure 1.
This project provides a ROS package for localization of an autonomous vehicle. The localization is based on a map consisting of ORB features. The mapping and localization module is taken from the <a href="https://github.com/raulmur/ORB_SLAM2">ORB-SLAM2</a> implementation. Our project builds on top of the [ROS-enabled version](https://gitlab.tu-berlin.de/breakdowncookie/ORB_SLAM2). In this extensions the map of ORB-features be saved to the disk as a reference for future runs along the same track. At high speeds, a map-based localization, may be prefered over a full SLAM due to better localization accuracy and global positioning reference. The use-case for the presented system is a closed circuit racing scenario. In a first step, we create a map at low speeds with a full SLAM. In further runs at higher speeds, we localize on the previously saved map to increase the localization accuracy. The usage is demonstrated with an example from the KITTI dataset. The flow chart of the extended functionality is shown in Figure 1.

<p align="center">
<img src="doc/orbslam2_extension_diagram.png">
Expand Down

0 comments on commit 536e2d5

Please sign in to comment.