Terry Taewoong Um

Seoul National University, GS Caltex (South Korea), Korea Institute of Science and Technology

Papers

Total Citations

146

H-Index

About

Terry Taewoong Um is a leading roboticist whose work bridges the frontiers of motion planning and robot learning. His primary research focuses on developing algorithms for robots operating under complex physical constraints, particularly on curved, high-dimensional configuration spaces known as constraint manifolds. Um’s most significant contribution is the creation of the Tangent Bundle Rapidly Exploring Random Tree (TB-RRT) algorithm, which revolutionized constrained motion planning by constructing random trees on tangent bundle approximations of the manifold rather than on the manifold itself. This work, published in 2014, has garnered 95 citations and remains a foundational reference in the field. Building on this, his earlier Tangent Space RRT (TS-RRT) algorithm, with 32 citations, established the core idea of leveraging tangent spaces for planning under holonomic constraints. Um has also made notable strides in robot model learning, introducing Independent Joint Learning, a novel task-to-task transfer learning scheme that enables robots to efficiently learn dynamic models from data. His work is distinguished by its mathematical rigor and practical impact, offering elegant solutions to some of the most challenging problems in robotics—from manipulation to locomotion—and inspiring a new generation of researchers to think beyond Euclidean spaces.