Gregory Dudek

A cross-section of selected publications across my career. A more comprehensive and current list (with on-line versions) is the publication list by topic; see also the (less current) Mobile Robotics Lab publications page.

• One of the early applications of LLMs to robotics: we used a foundation model to create a photo-acquisition robot that adapts across domains via an LLM that produces domain-specific criteria and evaluation metrics. Followups call this a photobot, though earlier (pre-LLM) work introduced the more general "photo documentarian" idea (final ICRA 2023 version).
• Prediction of 5G traffic on edge compute via transfer learning with a technique we call Similarity-based Elastic Weight Consolidation — One for All: Traffic Prediction at Heterogeneous 5G Edge, IEEE Globecom 2021 (best-paper award, GlobeCom 2021).
• "Generating adversarial driving scenarios in high-fidelity simulators" uses adversarial methods to improve autonomous driving agents — an early example of robustifying agents by making the environment part of the learning cycle (ICRA 2019 paper).
• Efficient methods for drivers to meet one another or passengers — Fast and Efficient Rendezvous in Street Networks, IEEE/RSJ IROS 2016.
• Searching for novelty in video to navigate a swimming robot toward maximally informative observations — Autonomous Adaptive Underwater Exploration using Online Topic Modeling (journal version in Intl. Journal of Robotics Research, 2013).
• Human-robot interaction (HRI) and dialog, validating HRI methods (2011).
• Graphical state space programming (2009–2011): GSSP material and code — a framework for lightweight robot controllers.
• Visual barcodes as variable-resolution fiducial targets: Fourier tags and an enhanced payload variant (2011).
• Inferring the topology of sensor networks from sparse event information: ICRA 2006; see also our 2010 IEEE Transactions on Robotics article.
• Recommender-system work with a large user-generated tag base: WI 2003.
• Comparing PCA to landmark-based pose estimation: IJCAI 2003.
• Using visual landmarks for pose estimation: CVPR 2001.
• Reconstruction of depth images from intensity data: WACV 2002.
• Multi-robot collaboration for robust exploration: Annals of Mathematics and AI, 2001.
• Learning landmarks for pose estimation: ICCV 1999.
• Taxonomy for multi-agent robotics and survey: Robot Teams (2002).
• Global localization: "Localizing a Robot with Minimum Travel," SIAM J. Comput., April 1998.
• Localization from sonar models: compressed postscript.
• Just-in-time sensing from sonar and laser data: ICRA 1996.
• Classification and use of sonar landmarks: IROS 1997.
• Computing "shape texture" for recognition: compressed postscript.
• 2-D shape-from-darkness and hidden surface recovery: IROS 1995.
• 3-D shape-from-darkness with iterative carving: CVPR 1998.

• Interview — "Dudek on robots in depth" (robotics and marine robotics).
• The future of robotics according to Dudek — one-minute clip.
• Aqua robot "dancing" — Aqua robot dance.
• Short biography (French) for the OCTAS award — biography video.
• Three-minute introduction to Gregory Dudek and the lab — quick intro for undergraduates.
• Bravo 2011 honorees (including Greg Dudek) — 30-second highlight.
• TEDx talk on robotics and finding interesting things automatically — TEDx talk.
• Fully autonomous vision-based airplane flight (UAV) — Xu & Dudek.
• Learning how a carpet "feels" for a Roomba — unsupervised learning demo.
• Robotics research video for the G20/G8 summit — summit video.
• Hovering gait for the Aqua robot — hovering gait demo.