CV

Researcher · Mobile Robotics Engineer

Korea Electronics Technology Institute (KETI), Intelligent Robotics Research Center

Multi-robot fleet management · autonomous navigation · SDR/OTA · simulation · anomaly detection

• Multi-robot fleet management and navigation: Open-RMF-based supervisory control of heterogeneous robots — task allocation, collision avoidance, lane closure and real-time re-planning; ROS 2 service/action-based task hand-over; validated in a 30-robot simulation and deployed on 4 real robots; DDS/Zenoh communication tuning
• Middleware for heterogeneous system integration: MQTT–ROS 2 bridge between robot and elevator (100% conversion success, 30–60 ms latency) enabling multi-floor autonomous navigation; custom TCP/IP protocol between AGV and manipulator for manufacturing logistics
• Precision docking and operational efficiency: two-stage docking evolved from ArUco to 2D-LiDAR geometric markers to ICP registration, ±3–5 cm charging-dock accuracy on a real robot; simulation-based battery charging logic
• SDR OTA deployment framework: Docker/Git container-level deployment with common code + robot-profile (YAML) parameter separation, so Navigation, Patrol, Elevator and Docking modules are swapped and verified without code changes (ICCAS 2025)
• Multi-robot anomaly detection and observability: PySide6 fault-injection GUI and unified CSV logging; rule-based four-class anomaly detection with a web dashboard (ICCAS 2025), then Transformer self-supervised (RESTAD) learning-based detection (ICROS 2026, Aff-F1 0.94)
• Mobile-robot navigation pipeline: mapping–localization–planning–control pipeline assembled from open-source components on simulation and real robots (TETRA_S, Jackal); outdoor stack with FAST-LIVO2 mapping, grid_map traversability and Pure Pursuit–LOS tracking
• Quadruped and biped platform simulation: Unitree Go2 multi-floor (stairs, elevator) testbed across Gazebo Harmonic, Isaac Sim 5.0 and MuJoCo; locomotion policy porting with terrain-based switching; VLM-orchestrated autonomous mission agent

Intern

Korea Institute of Robotics and Technology Convergence (KIRO), Robot Vocational Innovation Center

Integrated training environment for heterogeneous industrial and collaborative robots

• Analyzed the control systems of 8+ robots (Hyundai, FANUC, ABB, KUKA industrial arms; UR, Doosan, Neuromeka cobots), matched coordinate frames and communication specs per robot, and designed beginner/intermediate/advanced curricula and hands-on manuals
• Ran the center’s opening-ceremony demonstration with dozens of machines operating simultaneously without interference; designed cobot end-effector motions for stamping and signature events

B.S.

Keimyung University

Robotics Engineering

• Capstone design award — four-legged wheel-leg hybrid robot (Mechathon)
• Undergraduate researcher, Robot Dynamics and Intelligent Control Lab (2021.03 – 2022.01) — multi-DOF control system design, MFC-based actuator control

A Learning-based Anomaly Detection Framework for Multi-Robot Systems (first author)

Conference of the Institute of Control, Robotics and Systems (ICROS 2026)

Transformer self-supervised (RESTAD) multi-robot time-series anomaly detection, region-level Aff-F1 0.94

Anomaly Detection and Visualization Framework for Multi-Robot Systems Using RMF (first author)

ICCAS 2025, pp. 808–811 (IEEE Xplore)

Rule-based detection of four robot- and fleet-level anomaly types with a real-time web dashboard, validated with 30 simulated robots

Design and Implementation of an OTA Update Framework for Software-Defined Robots using ROS 2 and Docker (second author)

ICCAS 2025, pp. 1994–1997 (IEEE Xplore)

ROS 2 + Docker OTA framework for SDR with a public/on-premise dual-server architecture, proof-of-concept validation

Traversability Assessment and Path Planning Using Obstacle Recognition Information for Multi-Robot System (first author)

ICCAS 2024, pp. 1670–1671

Proactive re-planning by deactivating edges whose obstacle count exceeds a threshold; cooperative vs. non-cooperative comparison with four robots

Method for Cross-Utilization of Obstacle Recognition Information in a Multi-Robot System for Collaborative Path Planning (first author, presenter)

Conference of the Institute of Control, Robotics and Systems (ICROS 2024)

Obstacles from 2D LiDAR and RGB-D (YOLOv8) transformed into the global map frame and shared for collaborative path planning

Cooperative Autonomous Driving System for Multi-Mobile Robots Using Open-RMF (first author)

19th Korea Robotics Society Annual Conference (KROC 2024), pp. 455–456

Open-RMF + ROS Navigation server-client fleet management, real-robot cooperative driving validation

Cooperative mapping, environment recognition and autonomous driving for multiple mobile robots in large-scale indoor workspaces

National R&D project. Open-RMF-based cooperative multi-robot navigation; 30-robot simulation, 4 real TETRA_S/M robots sim-to-real.

• Open-RMF fleet and task system, namespace/TF design, sensor-fusion obstacle recognition and lane closure, MAPF analysis, fault injection and anomaly detection, yearly integration tests and certified testing

Business model development and robot implementation/validation for Software-Defined Robot (SDR) domain services

National R&D project. Elevator riding for multi-floor navigation, LiDAR-marker precision docking, GitHub + Docker OTA zero-downtime deployment, TRON1 biped and Go2 quadruped extensions.

• MQTT–ROS 2 bridge and NavigateMultiFloor action, ICP docking ±3–5 cm, OTA robot-profile structure, Go2 three-simulator multi-floor testbed, VLM autonomous mission agent

Collaborative-intelligence-based Robot Plus competitiveness support (manufacturing logistics automation)

National R&D project. Heterogeneous logistics robots (TETRA carts, stacker AMR, MiR) integrated under Open-RMF, AGV–manipulator TCP/IP protocol, docking control hand-over, PySide6 GUI.

• free_fleet modification and map alignment, stacker/ACS communication protocol design, YOLOv8 obstacle recognition demonstration, Isaac Sim integration study