ujinf.net
Ujin Kwon · 권우진

Applied algorithms, vehicle software, telemetry runtimes, and analysis tools

Mechanical & Computer Engineering student building nonlinear solvers, real-vehicle autonomy integrations, telemetry runtimes, and analysis workflows for vehicle dynamics, robotics, and autonomous systems.

Interests: Optimal Control · Numerical Optimization · Dynamics · Vehicle Software · Real-time Systems

Read by capability Browse projects

Selected Work

  • ballistic-solver: C++/Python intercept solver with drag dynamics, auxiliary residual correction, benchmarks, and an ICROS 2026 manuscript.
  • Mapless Autonomous Parking: Hyundai Ioniq parking system using FAST-LIO odometry, C++ LiDAR occupancy-grid mapping, delay-aware trajectory following, and CAN-facing control.
  • Racing Telemetry Stack: Raspberry Pi GNSS/RTK runtime with receiver configuration, buffered logging, remote monitoring, and recovery handling.
  • Racing Analyze GUI: MATLAB telemetry workbench for multi-run loading, segment comparison, replay, and metrics extraction.

Focus

Applied Algorithms

Nonlinear dynamics, residual transformation, damping, convergence handling, and explicit failure states.

Vehicle Autonomy

FAST-LIO odometry, C++ occupancy-grid mapping, response-delay-aware tracking, reverse-capable execution, and Ioniq CAN command integration.

Telemetry Operations

GNSS/RTK collection, browser monitoring, logging pipelines, and track-side recovery under imperfect conditions.

Analysis Workflows

Replay, segment metrics, multi-run comparison, and domain-specific inspection tools for driver review.

Project Highlights

Deployable Numerics

ballistic-solver shows a paper-backed residual method, solver design, API boundaries, packaging, diagnostics, and benchmark numbers.

View details
Planning + Integration

Mapless Autonomous Parking shows real-vehicle Ioniq validation, ROS 2/Autoware integration, native occupancy mapping, measured-delay control, and CAN-facing execution.

View details
Field Runtime

Racing Telemetry Stack shows real-device handling, correction ingestion, fault tolerance, browser monitoring, and operational data flow design.

View details
Analysis Workflow

Racing Analyze GUI provides a stateful analysis interface with replay, segment metrics, and pane-based telemetry inspection.

View details

Highlights

  • Award: KSAE 2024 Smart e-Mobility Competition (EV Division) — Encouragement Prize.
  • Research: autonomous driving team lead at HuVILab, a university vehicle-intelligence lab, focused on real-vehicle mapless parking, planning/control under constraints, and coaching systems.
  • Operations: telemetry runtime / monitoring / analysis collaboration with Luxon Racing Team in the GTA class of the O-NE SUPERRACE CHAMPIONSHIP.