Deming Wang

UniLaSalle Amiens (ESIEE-Amiens), Tongji University

Papers

Total Citations

H-Index

About

Deming Wang’s research lies at the intersection of robotics, computer vision, and intelligent perception, with a focus on enabling precise and safe robotic manipulation in complex environments. In his foundational work, “Optimal Trajectory Planning of Manipulators With Collision Detection and Avoidance” (1992, 44 citations), Wang introduced a pioneering method for planning collision-free paths by representing obstacles and robot segments as convex polyhedra, laying early groundwork for real-time motion planning. More recently, he has advanced 6-D object pose estimation—a critical capability for robotic grasping—by developing a multiscale point cloud transformer that leverages depth geometry to overcome the limitations of RGB-only methods (2022, 15 citations). His innovative “Three-Filters-to-Normal+” approach (2024) further refines depth-to-normal translation by revisiting discontinuity discrimination, enhancing surface normal estimation from depth images. Across his career, Wang has consistently bridged theoretical algorithms with practical vision-based measurement, contributing to safer, more accurate autonomous systems. His work continues to influence both industrial robotics and academic research in 3D perception and motion planning.