RESEARCH

Control, intelligence, and robotic systems.

We develop methods that connect dynamics, optimization, learning, and interaction—and validate them on robots operating beyond controlled demos.

Reactive Control research previewWatch experiment ↗

01 / 11

Reactive Control

Mobile manipulators in changing environments

We develop whole-body controllers that can insert, remove, and reprioritize tasks online, allowing mobile manipulators to respond smoothly to unexpected events.

  • Hierarchical QP
  • Task transition
  • Collision avoidance
Whole-body Control research previewWatch experiment ↗

02 / 11

Whole-body Control

Dynamic control for legged robots

Our work combines inverse dynamics, compliant control, trajectory optimization, and learned guidance for stable multi-contact locomotion and manipulation.

  • Inverse dynamics
  • Humanoids
  • Legged manipulation
Multi-Robot & HRI research previewWatch experiment ↗

03 / 11

Multi-Robot & HRI

Robots that collaborate with people and each other

We study coordinated manipulation and intuitive interfaces for teams of robots and humans performing physically demanding tasks together.

  • Cooperative transport
  • AR interface
  • Shared autonomy
AI-based Applications research previewWatch experiment ↗

04 / 11

AI-based Applications

Learning-enabled robotic systems

We connect perception and learning with control for applications ranging from dexterous teleoperation to automated scientific experiments.

  • Imitation learning
  • VLA
  • Robot learning
2.5D Autonomous Navigation research previewWatch experiment ↗

05 / 11

2.5D Autonomous Navigation

Mobility beyond flat, structured spaces

We combine traversability estimation, mapping, and planning so mobile robots can operate reliably in rough and unstructured environments.

  • 3D SLAM
  • Traversability
  • Rough terrain
Optimal Control research previewWatch experiment ↗

06 / 11

Optimal Control

Real-time decisions under constraints

We investigate DDP, MPPI, and constrained sampling-based control for fast, robust motion generation across robotic platforms.

  • MPPI
  • DDP
  • Trajectory optimization
Humanoid Control research previewWatch experiment ↗

07 / 11

Humanoid Control

Locomotion and loco-manipulation

Our humanoid research targets whole-body behaviors that integrate balance, locomotion, manipulation, and AI-based planning.

  • Locomotion
  • Loco-manipulation
  • Physical AI
Multi-Robot Control research previewWatch experiment ↗

08 / 11

Multi-Robot Control

Coordinated manipulation at scale

We develop planning and control methods for dual-arm, triple-arm, and heterogeneous robot systems working in tightly coupled tasks.

  • Dual-arm
  • Multi-arm
  • Coordination
Constrained Motion Planning research previewWatch experiment ↗

09 / 11

Constrained Motion Planning

Feasible planning on constraint manifolds

We develop geometry-aware sampling-based planners that satisfy equality and inequality constraints while efficiently exploring complex feasible motion spaces.

  • Riemannian metrics
  • Constraint manifolds
  • RRT
Visual Language Action research previewWatch experiment ↗

10 / 11

Visual Language Action

Language-conditioned contact-rich manipulation

We combine vision-language-action policies with force-informed residual learning to make contact-rich manipulation more adaptive and robust.

  • Vision-language-action
  • Residual RL
  • Contact-rich manipulation
Visual Language Navigation research previewExplore project ↗

11 / 11

Visual Language Navigation

Instance-enriched semantic navigation

We build instance-enriched semantic maps that connect open-vocabulary perception, room understanding, and object attributes for language-guided navigation.

  • Open-vocabulary mapping
  • Semantic navigation
  • Embodied AI

INFRASTRUCTURE

Platforms for ambitious experiments.

Facilities at Kyung Hee University and the Advanced Institute of Convergence Technology.

01

Unitree H1-2 full-size humanoid

02

Robros IGRIS-C full-size humanoid

03

Scout mobile manipulator with Franka arm

04

Hunter, Limo, TurtleBot 3/4 mobile robots

05

Go2, Go2-W, and A2 quadruped robots

06

FR5, Piper, and SO-ARM100 manipulators

07

Nokov Pluto motion-capture system

08

HoloLens 2, Apple Vision Pro, and Meta Quest 3

09

A6000-class GPU servers and workstations