Project: 3D Mapping, Localization and SLAM


STAR Center robot for 3D SLAM

In a collaboration between Prof. Ma Yi and Prof. Sören Schwertfeger this project aims to explore novel algorithms for robotic mapping, robot localization and the combined problem "Simultaneous Localization and mapping" (SLAM).

Towards this end we had our students collect more than 700 high-resolution, 360-degree RGB-D scans (those collect the color as well as the distance from the sensor for each pixel) of the SIST building. Additionally several high-resolution images where taken at each scan location for visual SLAM. See the video below for a small fraction of this data, depicting the lobby of the SIST building. In the future we plan to release this dataset to the public.

Additionally we are working on robotic mapping, localization and SLAM. Towards this end our first mapping robot is a Clearpath Jackal robot with a big sensor set, depicted on the right. It features two Velodyne HDL32e 3D laser scanners, two 2D laser scanners, three high-resolution cameras, a RGB-D camera, an advanced IMU and plenty of computation power. We are using it collect data for 3D SLAM and also for augmenting the data collected with the high-resolution scanner mentioned above.

References

[ibib]1[/ibib]
[ibib]53[/ibib]

 


Video showing a 3D map of the SIST lobby.

 


All electric car owned by SIST that we will use for research on
large-scale 3D SLAM and maybe autonomous diving in the future.