Research Projects
Vehicle Adaptation for End-to-end Autonomous Driving using Physics Informed Neural Network, Aug 2024 - Current
- Advisor: Jaeheung Park
- Robust vehicle adaptation of an end-to-end autonomous driving AI model
- Physical dimensions are parameterized to predict corresponding control output for various vehicle models
South Korean Ministry of Trade, Industry and Energy - Alchemist Project: Next Generation Humanoid, Jul 2024 - Current
- Advisor: Jaeheung Park
- Contribution to LLM-Humanoid connection for human-robot interaction
Alley 3D reconstruction AI model for fire trucks (Dudaji Start-up Company), Apr 2023 - Jul 2023
- Developed an AI model that accurately detects and reconstructs 3D objects in alleys, providing real-time feedback to fire truck operators
- The AI model determines whether fire trucks can safely navigate through the alleyways or not
- Integrated the model with a mapping server, allowing it to create blocker events and share them in real-time
Metal-Organic Frameworks Feature Regression, Sep 2022 - Dec 2022
- Advisor: Seungchul Lee
- Development of a regression model for inverse design prediction of MOF properties
- Automatic hyperparameter tuning for optimized regression model performance
- Project Repository Link
Autonomous Recycle Robot (Dudaji Company Internship), Jul 2022 ~ Aug 2022
- 6-joints-robot that classifies and does pick-and-place trashes by using ROS, Deep Learning, and computer vision
- Contributed to robot detection by camera, robot movement optimization via ROS, pick-and-place logic development, converting algorithm from computer vision to robot coordinate, and collecting plastic waste image data for classification
Autonomous Delivery Mobile Robot: Duckpod ver. 3 (SNU ARI Lab Internship), Dec 2021 ~ Feb 2022
- Advisor: Sungwoo Kim
- Contribution to the autonomy simulation and optimization of robot mobility for a 6-wheel, skid steering autonomous delivery robot on SNU campus
- Project Repository Link
Autonomous Parking Simulation, Sep 2021 ~ Dec 2021
- Advisor: Suhee Han
- Development of an autonomous parking simulator using Unity ML-Agent and reinforcement learning techniques
- Trained and evaluated the model with 8 million episodes for front, parallel, and diagonal parking scenarios, respectively
- Project Repository Link
- Program Download
Autonomous Drone, Feb 2021 ~ Jun 2021
- Advisor: Suhee Han
- Design and implementation of an autonomous drone system using ultrasonic sensors to avoid obstacles and navigate terrain safely
- Rule-based approach utilized for extreme environment exploration based on obstacle conditions
Auto Training Routine Recording System, Sep 2020 ~ Dec 2020
- Advisor: Chulhong Kim
- Development of a computer vision-based system to detect human pose and automatically record exercise routines for weight training
Auto Strike Zone Detection, Feb 2018 - Jun 2018
- Advisor: Sungjun Jung
- Development of an automated system for real-time detection of strike zones in baseball