Papers
7
Total Citations
131
H-Index
6
About
Michael J. Procopio is a machine learning researcher whose work centers on autonomous robot navigation, terrain segmentation, and adaptive learning systems for unstructured outdoor environments. His research addresses one of robotics' most persistent challenges: enabling robots to reliably identify safe, traversable paths in complex, real-world settings where conditions shift unpredictably over time. Procopio's most significant contribution lies in applying classifier ensembles and ensemble selection methods to terrain segmentation, combining stereo vision with machine learning to dramatically improve navigation robustness. His most-cited paper, "Learning Terrain Segmentation with Classifier Ensembles for Autonomous Robot Navigation in Unstructured Environments" (2009), has garnered 64 citations and remains a foundational reference in the field. Complementing this, his investigations into concept drift — how learned models degrade as environmental conditions evolve — produced influential work on long-term learning, dynamic ensemble adaptation, and online mixtures of experts, collectively cited dozens of times across the robotics and machine learning communities. A particularly practical thread in his research addresses imbalanced training data, a frequently overlooked obstacle in real-world terrain prediction. Across roughly half a dozen publications spanning 2007 to 2010, Procopio built a coherent, progressive research agenda that meaningfully advanced the state of machine learning-driven autonomous navigation.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4An experimental analysis of classifier ensembles for learning drifting concepts over time in autonomous outdoor robot navigation13 citations · 2007
- 5
- 6
- 7