Our research spans the full spectrum of autonomous robotics, from physical platforms to AI foundation models.

Division 1: Adaptive Robot Platform

Research Leader: Prof. Ji-Hwan Yoo (KAIST)

Co-Leader: Prof. Kee-Wook Kyung (KAIST)

We develop physical platforms that understand and adapt to real-world interactions through advanced sensing, actuation, and control systems.

Unstructured Environment Response Platform

  • Adaptive Actuators — Instantaneous response mechanisms for unpredictable environments
  • Multi-modal Tactile Sensors — High-resolution sensing for material, force, and slip detection
  • Intelligence-Embedded Modules — Distributed processing for rapid response
  • Multi-Robot Platforms — Coordinated systems for complex tasks

Human-Level Precision Interaction Interface

  • Dexterous Manipulation — Human-hand-level sensing and multi-DOF interfaces
  • Haptic Feedback Systems — Real-time force and tactile feedback
  • HRI-Based Collaborative Manipulation — Human-robot collaboration technologies

Real-Time Reflex Control System

  • Material Deformation Detection — Instant sensing of physical changes
  • Local Reflex Loops — Distributed control for immediate response
  • Distributed Self-Intelligence — Autonomous decision-making at each module

Division 2: 5D Robot AI

Research Leader: Prof. Gyu-Bin Lee (GIST)

Co-Leader: Prof. Sung-Eui Yoon (KAIST)

We create AI foundation models that enable robots to understand and predict physical interactions in complex real-world scenarios.

🌐

5D Scene Graph World Foundation Model

  • Autonomous Exploration — Self-directed data generation for 5D scene graphs
  • 5D World Foundation Model — Comprehensive understanding of object states and interactions
  • Physics-Informed Neural Networks (PINN) — Physics-embedded neural architectures
  • Multi-Robot Sensor Fusion — Heterogeneous data integration for ultra-realistic modeling
🤖

Whole-Body Locomotion and Manipulation Foundation Model

  • Multi-Modal Demonstration Learning — Training from diverse environment data
  • Integrated Locomotion-Manipulation Control — Unified whole-body control systems
  • Cross-Robot Transfer — Adaptation to various robot platforms
👁

Real-Time Robot Response Foundation Model

  • Multi-Modal Sensor Fusion — 5D world model learning systems
  • Situation Prediction Models — Anticipating environmental changes
  • Dynamic Object Avoidance — Real-time navigation in changing environments
  • Social Autonomous Navigation — Human-aware locomotion systems
Join Our Research Team