Teaching
2026 Spring 3 credits · Year 4

AI Aerospace Capstone Design 1

AI 우주항공 캡스톤 디자인 1

Team-based capstone project — reinforcement learning for rocket guidance, satellite data applications, computer vision, and AI agent systems in aerospace domains.

Capstone Aerospace Reinforcement Learning Satellite Computer Vision

Course Overview

  • Course Type: Capstone Design (Project-based)
  • Credits: 3 credits
  • Program: AI Aerospace (항공우주·미래모빌리티 융합전공)
  • Recommended Year: 4th year (Senior)
  • Class Size: 6 students
  • Instructors: Young Joon Lee (이영준), Yongjun Chang (장용준)

Course Description

A hands-on capstone design course where students form teams to develop AI applications in aerospace and related domains. Each team selects a research topic, conducts literature reviews, builds prototypes, and delivers final outputs including thesis writing and competition participation. Weekly full-group meetings ensure continuous progress and cross-team learning.

Learning Objectives

  1. Define and scope an AI research problem in an aerospace domain
  2. Conduct literature review and position work within existing research
  3. Design and implement AI solutions using appropriate frameworks and tools
  4. Present progress clearly and respond to technical feedback
  5. Write research papers suitable for conference or competition submission
  6. Collaborate with industry partners (e.g., Hanwha Systems)

Project Teams

Team A — Rocket Guidance Optimization

Reinforcement learning-based rocket guidance law optimization. Exploring RL algorithms for trajectory planning and control in launch vehicle systems.

Team B — Satellite & Environmental AI

Satellite data and computer vision applications — weather data (wind speed, solar radiation) for renewable energy site mapping, satellite attitude control with sensor/wheel design, and Jeju-specific tourism and running service applications using satellite imagery.

Individual Project — Election AI System

AI agent system for KBS election broadcast analysis — designing and building an autonomous analysis pipeline for the 2026 Korean local elections.

Course Structure

Phase 1 — Foundation (Weeks 1–4)

  • Team formation and topic selection
  • Literature review and research positioning
  • Technology stack definition and resource planning
  • Industry partnership exploration

Phase 2 — Development (Weeks 5–10)

  • Weekly progress presentations (individual, not team-based)
  • Data collection and preprocessing
  • Model design, training, and iteration
  • Guest lectures and external feedback sessions

Phase 3 — Delivery (Weeks 11–15)

  • System integration and evaluation
  • Thesis writing and revision
  • Competition submission preparation
  • Final presentations and demonstrations

Evaluation

  • Weekly Progress & Participation: 30%
  • Interim Presentations: 20%
  • Final Deliverable (Thesis + Demo): 40%
  • Peer Review: 10%

Tools & Platforms

  • Python, PyTorch, TensorFlow
  • AWS (cloud computing resources)
  • Satellite data APIs and GIS tools
  • Reinforcement learning frameworks (Stable Baselines, RLlib)
  • Computer vision libraries (OpenCV, YOLO)