Max Robotics — Global Service Robot Directory + US Certification

About

Dongyu Yan is a robotics and computer vision researcher whose work sits at the intersection of 3D scene reconstruction, neural implicit representations, and autonomous perception. His most recognized contribution, "Active Implicit Object Reconstruction Using Uncertainty-Guided Next-Best-View Optimization" (2023), has garnered 19 citations and addresses a critical challenge in autonomous mobile robotics: how to efficiently and accurately reconstruct objects by intelligently planning sensor viewpoints. By integrating emerging implicit neural representations with uncertainty-guided next-best-view optimization, Yan's approach strikes a principled balance between reconstruction accuracy and computational efficiency — a meaningful advance for real-world robotic deployment. Beyond active reconstruction, Yan has extended his expertise into panoramic perception and novel view synthesis. His work on MSI-NeRF explores the fusion of omni-depth information with generalizable Neural Radiance Fields using fisheye cameras, enabling six-degrees-of-freedom rendering for virtual reality and robot perception — pushing beyond the limitations of traditional panoramic methods. Together, these contributions reflect a coherent research vision centered on enabling machines to intelligently perceive and reconstruct their environments. Yan's work is increasingly relevant as autonomous systems demand richer, more efficient 3D understanding of the world around them.