I am a Ph.D. student at Kwangwoon University, Seoul, South Korea. My research focuses on learning-based decision-making, reliable sensing, and positioning for autonomous systems. Alongside academic research, I build AI-powered applications and research prototypes.
- Fault Diagnosis & Condition Monitoring: anomaly detection and reliability-aware analysis of sensor signals.
- Autonomous UAV Navigation: deep reinforcement learning for path planning and collision avoidance, including GPS-denied environments.
- Wireless & GNSS Positioning: RSS-based distance estimation, adaptive Kalman filtering, and robust positioning.
- Applied LLM Systems: on-device language models, retrieval-augmented generation, and practical AI software.
| Project | Overview |
|---|---|
| RSS-aware UAV Path Planning | Research implementation of UAV path planning and collision avoidance using RSS, EKF, and deep reinforcement learning. Paper |
| Deep Translation | Windows desktop translator powered by locally running LLMs, with GGUF model support and optional OpenAI-compatible servers. Latest release |
| Cell Decomposition-based UAV Path Planning | 3D UAV path planning with an A2C agent and cell decomposition, including a baseline comparison and pretrained models. Paper |
| DRL Path Planning Experiments | UAV navigation experiments with DQN and PPO; additional algorithms and multi-agent extensions are under development or archived. |
| Academic Writing RAG | Academic-writing assistant prototype exploring RAG, LangGraph, reviewer-response drafting, and reference formatting. |
First-author journal publications, most recent first.
-
AI-Driven Style Transfer Framework Based on 3D Gaussian Splatting for Immersive Experiences
K. Kim, B. Hwang, M. Lee, J. Kim, J. Seon, S. Kim, Y. Sun, S. Cho, J. Kim
Applied Sciences, 16(4), 1889 (2026). DOI -
Robust UAV Path Planning Using RSS in GPS-Denied and Dense Environments Based on Deep Reinforcement Learning
K. Kim, J. Seon, J. Kim, J. Kim, Y. Sun, S. Lee, S. Kim, B. Hwang, M. Lee, J. Kim
Electronics, 14(19), 3844 (2025). DOI · Code -
Enhanced 3D Outdoor Positioning Method Based on Adaptive Kalman Filter and Kernel Density Estimation for 6G Wireless System
K. Kim, S. Lee, B. Hwang, J. Kim, J. Seon, S. Kim, Y. Sun, J. Kim
Electronics, 13(23), 4623 (2024). DOI
- UAV Path Planning based on Deep Reinforcement Learning using Cell Decomposition Algorithm
K.-H. Kim, B. Hwang, J. Seon, S.-H. Kim, J.-Y. Kim
The Journal of the Institute of Internet, Broadcasting and Communication, 24(3), 15–20 (2024). DOI · Code
- Programming: Python, C/C++, C#, MATLAB, Bash
- Machine Learning & Simulation: PyTorch, deep reinforcement learning, Kalman filtering
- LLM Applications: RAG, LangChain, LangGraph, llama.cpp, .NET
- Development: Git, VS Code, CMake, Windows, Ubuntu/WSL

