Paper title
Mapping the Physical AI Landscape in Education
Target venue
한국피지컬AI학회 Journal, inaugural issue (June 2026)
Paper type
Systematic mapping study
Target length
6,000–9,000 words + database appendix
- Now → Apr 30 - Build survey database (target 150+ entries). Lock inclusion/exclusion criteria.
- May 1–15 - Run analysis: geographic distribution, hardware trends, curriculum maturity scoring.
- May 15–31 - Full draft written. Internal review.
- Early June - Submit to inaugural issue.
1 Introduction ~600 words
Why Physical AI in education, and why map it now? Establish the stakes: the field is moving fast, university curricula are lagging, and there is no consolidated reference for educators or policymakers building programs.
- Define “Physical AI” for the paper’s purposes (AI that acts on or through physical systems — distinct from pure software AI)
- Distinguish from adjacent terms: embodied AI, cyber-physical systems, robotics education, edge AI
- State the research questions: Where is Physical AI being taught? What hardware/software stacks dominate? What gaps exist? How does Korea compare globally?
- Brief note on your positionality as founder of the first Korean Physical AI academic society journal
2 Methodology ~800 words
Frame this explicitly as a systematic mapping study (cite Petersen et al. 2008 as your methodological anchor — it’s the canonical reference). This framing is more defensible than “literature review” and more appropriate for a field survey.
- Search strategy: Google Scholar, Scopus, IEEE Xplore, ACM DL, ASEE PEER, direct university catalog searches (Google: “physical AI course” site:edu, etc.)
- Search terms: “physical AI,” “embodied AI curriculum,” “edge AI education,” “robotics AI course,” “Jetson education,” “cyber-physical systems course”
- Inclusion criteria: University-level course or program, identifiable institution, taught 2018–2026, primary focus on AI + physical systems
- Exclusion criteria: K-12 only, purely theoretical (no hardware component), non-degree training programs
- Database schema: Country, Institution, Institution Type, Course Level, Delivery Format, Hardware Stack, Software Stack, Assessment Method, Year Introduced, Language, Open Resources (Y/N), Source URL
- Validation: Two-pass review; flag ambiguous entries; note entries confirmed vs. inferred
3 Results — Global Landscape ~1,800 words
Present the database findings. Use figures: a world map of course density, bar charts of hardware stack frequency, a timeline of course introductions by year.
- 3.1 Geographic distribution — where is Physical AI being taught? Likely heavy in US, EU, East Asia. Map it.
- 3.2 Institution type — R1 research universities vs. teaching-focused vs. polytechnics. Who is leading?
- 3.3 Curriculum maturity — dedicated programs vs. modules vs. elective courses. Propose a 3-level maturity model (Module → Course → Program).
- 3.4 Hardware ecosystems — NVIDIA Jetson, Raspberry Pi, Arduino, ROS-based systems. Frequency analysis.
- 3.5 Software stacks — PyTorch, TensorFlow, ROS2, OpenCV dominance. Which combinations appear together?
- 3.6 Assessment approaches — project-based vs. exam-based vs. publication-linked. Note that publication-linked assessment is rare and worth highlighting.
4 Results — Korean Context ~1,000 words
Your differentiator. No English-language paper has systematically characterized Korean Physical AI education. This section is why this paper belongs in the inaugural issue of a Korean Physical AI journal.
- 4.1 Current state of Physical AI in Korean universities — what exists, where, at what level
- 4.2 Government policy context — 과학기술정보통신부 AI policy, 교육부 SW/AI 교육 정책, how they shape curriculum
- 4.3 Infrastructure reality — budget constraints, lecturer-heavy staffing, lack of dedicated lab space. Honest characterization.
- 4.4 Comparison to global leaders — specific gaps and specific advantages (strong math/CS foundations, hardware manufacturing ecosystem)
- 4.5 Your own courses as mini-case studies — Autonomous Driving & ML, IoT, Python ML, Imaging-Based Medical Devices. Cite your GitHub Classroom papers here.
5 Discussion ~1,200 words
Interpret the findings. This is where your voice as a field-founder comes through.
- 5.1 Key gaps: Lack of open curriculum resources; hardware cost barriers; no standardized competency framework for Physical AI graduates
- 5.2 Emerging patterns: NVIDIA Jetson becoming the de facto platform; ROS2 adoption accelerating; simulation-first pedagogy growing
- 5.3 Recommendations for Korean institutions: prioritize module-level integration before full courses; leverage existing IoT/embedded infrastructure; use NVIDIA DLI to bridge training gaps
- 5.4 Propose a Physical AI Curriculum Framework: a 3×3 matrix of competency areas (Sensing / Inference / Actuation) × skill levels (Conceptual / Implementable / Publishable). This becomes a citable contribution that others reference.
- 5.5 Limitations: web-based search may miss unpublished courses; English-language search bias; rapidly changing field
6 Conclusion ~400 words
Summarize the map, restate the gaps, and signal the lab’s role in addressing them. End with a forward-looking statement about the inaugural journal’s role in building this field in Korea.
- Restate the 3 research questions and answer each in 1–2 sentences
- Propose the database as a living resource (link to public Google Sheet in appendix)
- Close with a call for curriculum sharing and open collaboration among Korean Physical AI researchers
+ Appendix — Course Database Public Google Sheet
Link to the full living database. Include a static snapshot table of the top 20–30 most complete entries in the paper body. The living spreadsheet becomes a citable research artifact on its own — other researchers will link to it, which builds your citation count over time.
Key citations to collect now
- Petersen et al. (2008) — “Systematic Mapping Studies in Software Engineering” — your methodological anchor
- Kitchenham & Charters (2007) — systematic review guidelines (for credibility)
- ROS2 / NVIDIA Jetson adoption papers from IEEE ICRA / IROS education tracks
- Korean government AI education policy documents (교육부, 과기정통부)
- Your own prior papers: GitHub Classroom (JPEE 2024/2025), AI in Education (JCCR 2025)
- Any ASEE papers on robotics/embedded AI curriculum (search ASEE PEER database)