SLAM & Semantic Mapping

My primary research area is SLAM and semantic mapping. I’ve developed a flexible mapping framework that supports a variety of sensors, feature types, and robot platforms, as showcased in the above video. An open source release is planned in Spring 2014.

Efficient Segmentation of Organized Point Clouds

Segmenting planes and clusters from RGB-D images is useful for many tasks, including semantic mapping, object tracking, and more. During my internship at Willow Garage, I developed a connected-component based approach to this, which is now available in the Point Cloud Library(PCL). You can find relevant papers describing the approach on my publications page.

Service Robot Interaction

I’m particularly interested in applying my perception and mapping work to service robots. The above is an example of how a user can interactively annotate object models and maps, for later use in service robotic tasks.