CV

Dong Hyeon Kim

Control & Robotics Engineer @ LUXROBO (2024.05 ~ Present)

Across changing domains, I turn engineering problems into mathematical models, implement them in algorithms and code, and validate them on real systems.

Robotics (Autonomous Driving, Humanoid)
Control Theory (PID, SMC, MPC)
State Estimation (Kalman Filter, FIR)
Multi-Agent Systems (Swarm Control, Cooperative Localization)
Sensor Fusion (LiDAR/UWB/Vision/IMU)
Perception (Object Detection, Tracking)

Work Experience

LUXROBO — Engineer 2024.05 ~ Present

Circuit Modeling & Simulation 2026.03 ~ Present

Board-level thermal simulation engine development (AI Team · new business initiative)

Developing a PCB physics simulation backend platform — a board-level thermal analysis engine driven by ECAD geometry and netlist data
Steady-state thermal analysis based on a 3D finite-volume method (FVM), modeling circuit physics such as component power estimation, current distribution, and IR drop
Building the Python/Rust engine and FastAPI backend API, validating against a commercial thermal solver (ANSYS Icepak)

PythonRustFastAPIFVMThermal Analysis

Autonomous Driving Module for Golf Carts 2025.06 ~ 2026.02

Development of vehicle localization and driving control modules

Implemented and validated a high-precision localization algorithm based on RTK-GPS
Developed path planning and steering control algorithms (e.g., Pure Pursuit)
Built a MATLAB-based simulation environment and conducted real-vehicle testing

Details

Built an offline MATLAB simulation environment ahead of real-vehicle testing, based on real driving logs (.bin)
Progressively validated an EKF prediction/correction structure based on the Bicycle Model
State vector: v, SlipAngle, YawRate, Yaw, x, y
Determined GPS measurement validity using a Chi-square Gate
Dynamically adjusted the R matrix based on GPS quality, with INS/GPS mode-switching logic
Compared localization results between simulation and firmware, with a real-time driving-log replay environment
Analyzed bottlenecks based on firmware TASK priority and execution performance

GNSSRTK-GPSCMATLAB AutonomousDrivingLocalizationSensorFusionEKFBicycleModelKalmanFilterVehicleDynamicsEmbeddedC

Smart Parking System 2024.05 ~ 2025.05 (Delivered)

Development of video processing server and lighting control server

Simultaneously controlled and analyzed about 30 CCTV cameras on a single server (RTSP stream processing); separated fisheye/one-way camera pipelines for scalability
Detected objects (vehicles/pedestrians) and analyzed speed using Optical Flow and YOLOv7 (ONNX), CUDA-accelerated
Designed a Redis-based event processing structure; automatic identification of nearby cameras via RSSI signal comparison
Real-time dynamic control of lighting devices over gRPC (integrated handling of CCTV, BLE, and emergency-bell events)

C++GoRedisgRPCOpenCVYOLOv7RTSP SmartParkingComputerVisionOpticalFlowMultiCameraProcessing

Research Experience

Advanced Control System Lab (Korea Univ.) — M.S. Research 2022.03 ~ 2024.02

Development of OAFMCL (Odometry Aided Finite-Memory Cooperative Localization) algorithm

Designed a Finite-Memory (FIR)-based cooperative localization algorithm, resolving the error-accumulation problem of recursive (IIR) estimation
Fused an external-sensor-based FMCL with an internal-sensor-based Odometry-NN via a neural network, achieving robustness to walking vibration in humanoid robots (using joint-motion data as input)
Reduced error by up to about 80% (vs. LHOL) compared to existing algorithms in Robot Kidnapping scenarios, with an average error of 0.07 m
Verified through demonstration experiments in a multi-humanoid-robot environment (RTLS Leader + LiDAR Followers)
Experimental results (RTAMSE, meter): Normal condition — OAFMCL 0.0474 vs MFDKF 0.1527 / NLSPGO 0.1012 / LHOL 0.0857; Robot Kidnapping — OAFMCL 0.0687 vs MFDKF 0.1103 / NLSPGO 0.0881 / LHOL 0.3791
Structure: the Leader carries an RTLS tag (absolute position from fixed anchors), Followers use low-cost LiDAR (relative distance/direction); FMCL + Odometry-NN → Fusion NN → final position estimate

National Research Foundation of Korea (NRF) Project 2022.03 ~ 2024.02

AI-based Cooperative Autonomous Swarm Anti-Drone System for Neutralizing Illegal Drones (Multi-UAV Formation Control)

Developed a Finite-Memory distributed controller and collision-avoidance algorithm for Multi-UAV Formation Control
Realized decentralized cooperative control without central communication using a graph-based interaction topology model and a Virtual Leader structure
Validated via MATLAB simulation and Pixhawk hardware experiments (Virtual Leader + 4 Followers), with formation error within ±0.2 m

PixhawkDistributedControlFormationControlMultiUAVVirtualLeaderGraphModelROS

Korea Institute of Science and Technology (KIST), Robot Vision Lab — Intern 2021.01 ~ 2021.02

Researched ROS-based image processing algorithms and tested their application to mobile robot navigation

ROBIT (Robot Game Team, Kwangwoon Univ.) — Member · Humanoid Team Leader (2020) 2019.01 ~ 2022.02

Led development of the full robot system (circuits, firmware, vision, control) for the RoboCup Humanoid League
Developed competition vision processing based on OpenCV (color-based techniques robust to lighting changes), improving indoor/outdoor recognition stability
Particle Filter-based visual localization: extracted soccer-field lines → converted to a likelihood model → applied Particle Filter weighting
Designed and built humanoid control circuits; developed STM-based firmware (motor/sensor control)
Participated in domestic and international competitions, including a team trip to RoboCup 2019 Sydney, earning multiple awards
Engaged in outreach activities such as operating robot demonstration booths at science festivals and exhibitions

HumanoidRobotComputerVisionVisualLocalizationParticleFilterSTM32EmbeddedSystem

Capstone Design — Contactless Caddy Service Robot — Team Leader 2021.01 ~ 2021.12

ROS-based autonomous caddy robot: 2D LiDAR obstacle avoidance + GPS user-following, outdoor prototype
Generated target points from the user's remote-control GPS position; avoided dynamic/static obstacles and followed the user via a Potential Field driving algorithm
Designed a Potential Field structure combining an attractive force toward the user and a repulsive force from obstacles
GPS/IMU/2D LiDAR perception pipeline; IMU-based heading estimation with north-referenced coordinate normalization
Validated in ODE simulation first, then applied stepwise to real-vehicle driving; directly designed and integrated the power/control circuits

AutonomousRobotPotentialFieldUserFollowingObstacleAvoidanceMobileRobot

Publications

D. H. Kim, J. M. Pak, P. Shi, and C. K. Ahn, “Finite-Memory CL Based on Multisensor Information Fusion Using Neural Networks for Multiple Humanoid Robots,” IEEE Transactions on Industrial Electronics, vol. 73, no. 1, pp. 1384–1393, Jan. 2026. DOI: 10.1109/TIE.2025.3600539 — IF 7.4 (2025 JCR), Q1
J. Y. Ko, J. M. Pak, D. H. Kim, and C. K. Ahn, “Secure Event-Triggered Distributed Finite-Memory Localization for Multiple Mobile Robots Under Cyberattacks,” IEEE Transactions on Industrial Electronics, vol. 71, no. 10, pp. 12931–12941, Oct. 2024. — IF 7.2 (2024 JCR), Q1
D. H. Kim, H. H. Kang, K. S. Kim, and C. K. Ahn, “3D Localization and Cyber-Attack Estimation of a Drone using RBFNN-based Finite Memory State Estimator,” Proceedings of the ICROS (Institute of Control, Robotics and Systems) Domestic Conference, pp. 294–295, 2023.

Education

Korea University 2022.03 ~ 2024.02

M.S. in Electrical Engineering

Advanced Control System Lab (Advisor: Prof. Choon Ki Ahn)

Fusion of Neural Networks and Finite-Memory Estimation for Distributed Localization of Multiple Humanoid Robots

Teaching Assistant — Engineering Mathematics (2022.03 ~ 2022.08), Control Engineering (2022.09 ~ 2023.02)

Kwangwoon University 2018.03 ~ 2022.02

B.S. in Robotics (Intelligent Systems Major)

Full-tuition scholarship for ROBIT (robot game team) activities (2019 ~ 2021)

Honors & Awards

Grand Prize — KATECH Future Mobility Idea Competition (Korea Automotive Technology Institute) 2023.09
Excellence Award (University President's Award) — 17th Kwangwoon ICT Exhibition (KWIX) 2021.09
Grand Prize (Presidential Award) — International Robot Contest 2020, Humanoid Robot Sports (hosted by the Ministry of Trade, Industry and Energy & Seoul National University of Science and Technology) 2020.11
3rd Place — International Robot Contest 2020, Humanoid Robot Sports — Intelligent High-Tech League 2020.11
Champion — RoboCup 2020 Korea Open Humanoid Soccer Kid 2020.08
2nd Place — RoboCup 2019 Sydney, Humanoid League TeenSize Drop-In Challenge 2019.07
3rd Place — RoboCup 2019 Sydney, Humanoid League TeenSize Technical Challenge 2019.07
Champion — RoboCup 2019 Korea Open Humanoid Soccer Kid 2019.02

Grand Prize, KATECH Future Mobility Idea Competition (2023)

Excellence Award (University President's Award), 17th Kwangwoon ICT Exhibition (KWIX) (2021)

Grand Prize (Presidential Award), International Robot Contest 2020, Humanoid Robot Sports (2020)

3rd Place, International Robot Contest 2020, Humanoid Robot Sports — Intelligent High-Tech League (2020)

Champion, RoboCup 2020 Korea Open Humanoid Soccer Kid (2020)

2nd Place, RoboCup 2019 Sydney, Humanoid League TeenSize Drop-In Challenge (2019)

3rd Place, RoboCup 2019 Sydney, Humanoid League TeenSize Technical Challenge (2019)

Champion, RoboCup 2019 Korea Open Humanoid Soccer Kid (2019)

Skills

Control & Estimation Kalman Filter (KF/EKF)Finite-Memory Filter (FIR)Particle FilterSensor FusionPIDPure PursuitPotential Field

Modeling & Simulation MATLABFVMThermal Analysis

Computer Vision OpenCVYOLOv7CUDARTSPCamera

Embedded Systems CSTM32PixhawkROS/ROS2RTK-GPS/GNSSLiDARIMUUWB

Backend & Infrastructure C++PythonGo (Golang)RedisgRPCDockerGit

Language Proficiency TOEIC Speaking 140 — Intermediate High (2024)

김동현 (Dong Hyeon Kim)

Control & Robotics Engineer @ LUXROBO (2024.05 ~ 현재)

도메인이 바뀌어도, 공학 문제를 수학 모델로 만들고 알고리즘과 코드로 구현해 실제 시스템에서 검증.

Robotics (Autonomous Driving, Humanoid)
Control Theory (PID, SMC, MPC)
State Estimation (Kalman Filter, FIR)
Multi-Agent Systems (Swarm Control, Cooperative Localization)
Sensor Fusion (LiDAR/UWB/Vision/IMU)
Perception (Object Detection, Tracking)

Work Experience

LUXROBO — Engineer 2024.05 ~ 현재

회로 모델링 & 시뮬레이션 개발 2026.03 ~ 현재

보드 레벨 열해석 시뮬레이션 엔진 개발 (AI팀 · 회사 신규 사업)

PCB 물리 시뮬레이션 백엔드 플랫폼 개발 중 — ECAD 기하·네트리스트 기반 보드 레벨 열해석 엔진
3D 유한체적법(FVM) 기반 정상상태 열해석과 부품 발열원 추정·전류 분포·IR drop 등 회로 물리 모델링
Python·Rust 엔진 및 FastAPI 백엔드 API 개발 중, 상용 열해석 도구(ANSYS Icepak) 대비 검증 중

PythonRustFastAPIFVMThermal Analysis

골프카트 자율주행 모듈 개발 2025.06 ~ 2026.02

차량 위치 추정 및 주행 제어 모듈 개발

RTK-GPS 기반 고정밀 위치 추정 알고리즘 구현 및 검증
주행 경로 계획 및 조향 제어 알고리즘(Pure Pursuit 등) 개발
MATLAB 기반 시뮬레이션 환경 구축 및 실차 테스트 수행

Details

실차 주행 로그(.bin) 기반 MATLAB 오프라인 시뮬레이션 선행 구축
자전거 모델(Bicycle Model) 기반 EKF 예측·보정 구조 단계적 검증
상태 벡터 v, SlipAngle, YawRate, Yaw, x, y
Chi-square Gate 기반 GPS 측정값 유효성 판별
GPS 품질에 따른 R 행렬 동적 조정 및 INS/GPS 모드 전환 로직
시뮬레이션-펌웨어 간 위치 추정 결과 비교 및 실시간 주행 로그 재생 환경
펌웨어 TASK 우선순위·실행 성능 기반 병목 분석

GNSSRTK-GPSCMATLAB AutonomousDrivingLocalizationSensorFusionEKFBicycleModelKalmanFilterVehicleDynamicsEmbeddedC

스마트 주차 시스템 개발 2024.05 ~ 2025.05 (납품 완료)

영상 처리 서버 및 조명 제어 서버 개발

단일 서버에서 약 30대 CCTV 카메라 동시 제어 및 분석 (RTSP 스트림 처리), 어안(Fisheye)/단방향(One-way) 카메라 파이프라인 분리로 확장성 확보
Optical Flow 및 YOLOv7(ONNX) 기반 객체(차량/보행자) 탐지 및 속도 분석 (CUDA 가속)
Redis 기반 이벤트 처리 구조 설계, RSSI 신호 비교 기반 근접 카메라 자동 식별
gRPC 통신 기반 실시간 조명 장치 동적 제어 (CCTV·BLE·비상벨 이벤트 통합 처리)

C++GoRedisgRPCOpenCVYOLOv7RTSP SmartParkingComputerVisionOpticalFlowMultiCameraProcessing

Research Experience

Advanced Control System Lab (Korea Univ.) — 석사 연구 2022.03 ~ 2024.02

OAFMCL (Odometry Aided Finite-Memory Cooperative Localization) 알고리즘 개발

유한 메모리(FIR) 기반 협력 위치 추정 알고리즘 설계로 재귀 추정(IIR)의 오차 누적 문제 해결
외부 센서 기반 FMCL과 내부 센서 기반 Odometry-NN을 신경망으로 융합, 로봇 보행 진동에 강인성 확보 (관절 움직임 데이터를 입력으로 사용)
Robot Kidnapping 상황에서 기존 알고리즘 대비 최대 약 80%(LHOL 대비) 오차 감소, 평균 오차 0.07m
다중 휴머노이드 로봇 환경(RTLS Leader + LiDAR Followers)에서 실증 실험
실험 수치 (RTAMSE, meter): Normal condition — OAFMCL 0.0474 vs MFDKF 0.1527 / NLSPGO 0.1012 / LHOL 0.0857; Robot Kidnapping — OAFMCL 0.0687 vs MFDKF 0.1103 / NLSPGO 0.0881 / LHOL 0.3791
구조: Leader는 RTLS 태그(고정 앵커 절대 위치), Follower는 저가형 LiDAR(상대 거리·방향); FMCL + Odometry-NN → Fusion NN → 최종 위치 추정

한국연구재단(NRF) 프로젝트 2022.03 ~ 2024.02

불법드론 무력화를 위한 AI 기반 협력적 자율주행 스웜 안티드론 시스템 (Multi-UAV Formation Control)

Multi-UAV Formation Control을 위한 Finite-Memory 분산 제어기 및 충돌 회피 알고리즘 개발
그래프 기반 상호위상 모델과 Virtual Leader 구조로 중앙 통신 없는 분산 협력 제어 실현
MATLAB 시뮬레이션 및 Pixhawk 하드웨어(Virtual Leader + 팔로워 4대) 실험, 포메이션 오차 ±0.2m 이내 검증

PixhawkDistributedControlFormationControlMultiUAVVirtualLeaderGraphModelROS

한국과학기술연구원(KIST) Robot Vision Lab — 인턴 2021.01 ~ 2021.02

ROS 기반 영상처리 알고리즘 연구 및 모바일 로봇 주행 적용 테스트

광운대학교 로봇게임단 ROBIT — 단원 · Humanoid Team Leader (2020) 2019.01 ~ 2022.02

RoboCup 휴머노이드 리그 참가용 로봇 전체 시스템(회로, 펌웨어, 비전, 제어) 개발 주도
OpenCV 기반 대회용 영상 처리(조명 변화에 강인한 색상 기반 기법), 실내·실외 인식 안정성 향상
Particle Filter 기반 영상 위치 추정(Visual Localization): 축구장 라인 추출 → Likelihood 모델 변환 → Particle Filter 가중치 활용
휴머노이드 제어 회로 설계·제작, STM 기반 펌웨어 개발 (모터·센서 제어)
RoboCup 2019 Sydney 국제대회 원정 등 국내외 대회 참가 및 다수 입상
과학축제·전시 부스 운영 등 로봇 시연 대외 활동

HumanoidRobotComputerVisionVisualLocalizationParticleFilterSTM32EmbeddedSystem

Capstone Design — 비대면 캐디 서비스 로봇 — 팀 리더 2021.01 ~ 2021.12

ROS 기반 자율주행 캐디 로봇: 2D LiDAR 장애물 회피 + GPS 사용자 추종, 실외 환경 프로토타입
사용자 리모컨 GPS 위치로 목표점 생성, Potential Field 주행 알고리즘으로 동적·정적 장애물 회피 및 사용자 추종
사용자 방향 인력 + 장애물 척력 결합 Potential Field 구조 설계
GPS·IMU·2D LiDAR 인지 파이프라인, IMU 기반 heading 추정과 북쪽 기준 좌표계 정규화
ODE 시뮬레이션 선행 검증 후 실차 주행 단계적 적용; 전원·제어 회로 직접 설계·통합

AutonomousRobotPotentialFieldUserFollowingObstacleAvoidanceMobileRobot

Publications

D. H. Kim, J. M. Pak, P. Shi, and C. K. Ahn, “Finite-Memory CL Based on Multisensor Information Fusion Using Neural Networks for Multiple Humanoid Robots,” IEEE Transactions on Industrial Electronics, vol. 73, no. 1, pp. 1384–1393, Jan. 2026. DOI: 10.1109/TIE.2025.3600539 — IF 7.4 (2025 JCR), Q1
J. Y. Ko, J. M. Pak, D. H. Kim, and C. K. Ahn, “Secure Event-Triggered Distributed Finite-Memory Localization for Multiple Mobile Robots Under Cyberattacks,” IEEE Transactions on Industrial Electronics, vol. 71, no. 10, pp. 12931–12941, Oct. 2024. — IF 7.2 (2024 JCR), Q1
D. H. Kim, H. H. Kang, K. S. Kim, and C. K. Ahn, “3D Localization and Cyber-Attack Estimation of a Drone using RBFNN-based Finite Memory State Estimator,” 제어로봇시스템학회 국내학술대회 논문집, pp. 294–295, 2023.