Max Robotics — Global Service Robot Directory + US Certification

About

Edgar Sucar is a robotics and computer vision researcher whose work sits at the intersection of neural scene representations, real-time perception, and robot understanding of 3D environments. His research focuses on enabling robots to build rich, actionable models of the world around them — from reconstructing object geometry to understanding complex, cluttered scenes in real time. Sucar's most cited work, "iSDF" (2022, 125 citations), introduced a pioneering approach to real-time neural signed distance field optimization, allowing robots to continuously learn collision costs and spatial gradients online — a critical capability for safe navigation and manipulation. His influential "MoreFusion" papers (2020, 103 citations) tackled the challenging problem of 6D object pose estimation within multi-object scenes, combining volumetric fusion with object-level reasoning to handle occlusion and contact. His NodeSLAM work further advanced learned object descriptors for multi-view shape reconstruction, bridging neural representations with probabilistic rendering. More recently, Sucar has pushed toward open-set scene understanding, fusing pretrained feature networks with neural scene representations to recognize novel, previously unseen objects — a significant step toward generalizable robot perception. Across his body of work, Sucar consistently bridges theoretical elegance with real-world robotic applicability.