Delin Feng

Papers

Total Citations

H-Index

About

Delin Feng is a robotics researcher whose work centers on advancing spatial intelligence for mobile robots, particularly in the areas of localization, mapping, and semantic understanding. Feng’s major contributions include the development of **FloorplanNet**, a learning-based method that enables robots to match their environmental observations with building floorplans using geometric, semantic, and topological cues—a task humans perform intuitively. This work, with 9 citations, bridges the gap between human spatial reasoning and robotic perception. Feng also introduced **osmAG**, a novel map format based on OpenStreetMap XML that stores hierarchical, topometric semantic maps for indoor and outdoor multi-floor environments, addressing a critical gap in robotics mapping standards (7 citations). Additionally, Feng built **Cluster on Wheels**, a compact 16-node computing cluster designed for a future robot to collect and process massive sensor data for SLAM research—the first known use of such a cluster in robotics. These innovations demonstrate Feng’s commitment to creating more intuitive, efficient, and scalable mapping and localization systems, with direct implications for autonomous navigation in complex real-world environments.