Alejandro Fontan

Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR), Queensland University of Technology

Papers

Total Citations

107

H-Index

About

Alejandro Fontan is a leading researcher in visual localization and autonomous navigation for planetary exploration, with a focus on making robotic systems more reliable and efficient in extreme environments. His work spans visual-inertial odometry, direct RGB-D SLAM, and scene coordinate regression, with a particular emphasis on information-theoretic approaches to sensor data selection. Fontan’s most cited work, the MADMAX dataset (54 citations), provides a critical benchmark for visual-inertial rover navigation on Mars, enabling the development of vision-based autonomy for planetary rovers under representative optical conditions. He introduced an information-driven point selection method for direct RGB-D odometry (33 citations), which reduces computational load while preserving accuracy—a key contribution for resource-constrained robotic platforms. More recently, Fontan proposed FocusTune (17 citations), a focus-guided sampling technique that improves visual localization by directing models toward geometrically critical regions for 3D point triangulation. His ongoing work includes adaptive outlier thresholding for bundle adjustment and forward prediction of localization failure, addressing the critical need for self-aware, safety-conscious navigation systems. Fontan’s research is shaping the future of autonomous exploration, from Mars rovers to terrestrial robotics.