Mercator Project: Mapping

The SBM Algorithm: Real-Time On-line Mapping and Localization

We demonstrate mapping with and without odometry. All results have been obtained in real-time. The SBM Algorithm combines fast sample-based (probabilistic) methods with Newton's search for density estimation and building the most likely map.

When the background flashed yellow, the map has been disturbed manually (30 degrees rotated, 50 cem shifted) to demonstrate the robustness of the approach

With Odometry: Animated GIF and Final Map.

Mapping Without Odometry. This experiment demonstrates that the SBM Algorithm can be applied to build maps with free-moving objects without any sensing of ego-motion (eg, people, gliders): Animated GIF and Final Map.

Another example for mapping without odometry: Animated GIF and Final Map.

Datasets: with odometry: and without odometry (click to enlarge)

Last Updated $Date: 98/11/24 10:46:11 $
Send questions or comments to da0g+@cs.cmu.edu