Rodolphe Dubois
Papers
4
Total Citations
44
H-Index
4
About
Rodolphe Dubois is a leading researcher in decentralized multi-robot perception, specializing in Simultaneous Localization And Mapping (SLAM) for collaborative robotic systems. His work addresses the critical challenge of enabling multiple robots—both ground and aerial—to jointly build accurate 3D maps without relying on a central server. Dubois’s major contributions include a pioneering dense decentralized SLAM algorithm that uses locally consistent Truncated Signed Distance Function (TSDF) submaps, allowing robots to share and merge 3D data efficiently. He also developed two innovative data-sharing strategies for visual-inertial SLAM—Factor Sparsification for Visual-Inertial Packets (FS-VIP) and Min-K-Cover Selection (MKCS-VIP)—which optimize communication bandwidth while preserving mapping accuracy. To support the community, Dubois introduced the AirMuseum dataset, a heterogeneous multi-robot benchmark featuring five indoor scenarios with stereo-visual and inertial data from both ground and aerial platforms. His most cited works have garnered over 44 citations, reflecting their impact on advancing collaborative autonomy. Dubois’s research is foundational for applications in search-and-rescue, industrial inspection, and autonomous exploration, where robust, decentralized mapping is essential.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4