cv

Research Intern (since Sep 2025) · M.S. Incoming (Spring 2026)

Seoul National University

Autonomous Robot Intelligence Lab (ARIL)

B.S.

Kyung Hee University

Electronic Engineering, Computer Engineering

Research Intern

Autonomous Driving Intelligence Lab (ETRI, Korea)

• Developed an auto-labeling tool that extracts lanes/trajectories/actors/work-zones from surround-view + LiDAR + HD maps, generating VQA fine-tuning JSONL.
• Used vision-fine-tuned InternVL 3.0 and GPT-4o/5 (API) to produce scene-understanding and action-decision outputs for end-to-end autonomous driving.

Autonomous Driving Developer

Immersion

• Led Perception & Navigation; built an end-to-end outdoor autonomy stack from hardware design and sensing to path planning and control.
• Designed and integrated a prediction-based avoidance module, improving dynamic obstacle handling in the existing autonomy stack.

Undergraduate Research Assistant

Human-Robot Interaction Lab, Kyung Hee University

Development of Autonomous Pick-and-Place Robot System with Multi-Floor Navigation

Journal of Korea Robotics Society

Under revision. Authors: Seongmin Hong, Yujin Jung, Yunjeong Choi, Seju Jang, Donghan Kim.

Academic Excellence Award (Top Rank in Department)

Kyung-Hee University, College of Electronics and Information

Grand Prize, Capstone Design Competition

LINC 3.0 Project Group, Kyung-Hee University

Excellence Award, Graduation Project

Kyung-Hee University, College of Electronics and Information

Encouragement Award, Autonomous Aviation Mini-drone Competition

Korean Institute of Electrical Engineers

2nd Prize, Autonomous Aviation Mini-drone Competition

Korean Institute of Electrical Engineers

Enhancing Diffusion-Based End-to-End Autonomous Driving with VLM

• Improved trajectory diversity and performance with VLM-guided dynamic anchor sampling; switched to a lightweight YOLO sampler for real-time with comparable results.
• Built a VLM-based Trajectory Scorer that re-ranks candidates by centerline alignment and collision-avoidance metrics, improving final path selection.

Autonomous Pick-and-Place Robot System with Multi-Floor Navigation (Husky + UR3)

• Developed an autonomous multi-floor pick-and-place robot system using a Husky-UR3.
• Applied hybrid maps, AI vision, and AR-marker to ensure safe navigation and accurate button press.

Outdoor Advertising Robot (Immersion Project)

• Enhanced dynamic obstacle avoidance by predicting obstacle motion using ByteTrack + Kalman filter.
• Developed a GUI-linked behavior tree for flexible and user-driven navigation control.

Woowa Bros. Robot Delivery Challenge

• Stabilized GPS-denied localization by fusing 3D LiDAR-based pose and IMU on a prebuilt 3D map.
• Calculated pedestrian speed and direction from YOLOv5 and LiDAR data, integrating collision risk into DWA for safe local path planning.

Lane-Keeping System in CARLA Simulator

• Developed a robust lane-keeping system using YOLOPv2, BEV, and Kalman filtering.
• Applied quadratic curve fitting and Stanley controller to compute accurate steering angles.