RESEARCH
I recently defended my thesis proposal. Here are the
slides from my presentation and the written
thesis proposal document.
My dissertation research focuses on achieving long-term learning through ensembles of linear models. Currently, this research is being applied towards the unsolved problem of robot navigation in unstructured outdoor environments.
Publications
Michael J. Procopio, Gregory Grudic, and Jane Mulligan.
Long-Term Learning Using Multiple Models for Outdoor Autonomous Robot Navigation.
Submitted to IROS 2007, April 2007. (
PDF)
Michael J. Procopio, Thomas Strohmann, Adam R. Bates, Gregory Grudic, and Jane Mulligan.
Using Binary Classifiers to Augment Stereo Vision for Enhanced Autonomous Robot Navigation. University of Colorado at Boulder, Boulder, Colorado 80309-430, Tech. Rep. CU-CS-1027-07, April 2007. (
PDF)
John M. Linebarger, Andrew J. Scholand, Mark A. Ehlen, and Michael J. Procopio.
Benefits of Synchronous Collaboration Support for an Application-Centered Analysis Team Working on Complex Problems: a Case Study. In Proceedings of the 2005 international ACM SIGGROUP Conference on Supporting Group Work (Sanibel Island, Florida, USA, November 06 - 09, 2005). GROUP '05. (
PDF)
Michael J. Procopio.
YCab.NET: Decentralized Collaboration Groupware for Mobile Devices using the Microsoft .NET Framework. Master's of Science Thesis, University of Florida, 2002. (
PDF)
Michael J. Procopio.
An Experimental Analysis of the Simple Genetic Algorithm. Undergraduate High Honors Thesis, University of Florida, 2002. (
PDF)
Datasets
As part of my dissertation research, I continue to extract and hand-label datasets from the
DARPA LAGR program. Here they are converted from the proprietary LAGR format (which requires the LAGR robot APIs to read) and put in a more accessible MATLAB format.
Hand-Labeled Robot Datasets (IROS Paper)
Braitenberg Vehicles
Please see my research on
Braitenberg Vehicles.
Pixelwise Image Labeler
I've created an application to aid in the labeling of images in a dataset. See my page on
Pixelwise Image Labeler.