Coding Path C — AI Pose Rescue Game

Coding — Path C

AI Pose Rescue Game

Build a rescue mission game in MRT Scratch AI Studio — then train a Google Teachable Machine AI model so your body movements control the game. No keyboard needed.

🧠 Year 5 – Year 12 👤 Individual or Team (up to 5) 🎮 AI-Controlled Game

In March 2026, the Great Western Highway at Victoria Pass — the main road between Sydney and the Blue Mountains — was shut down in both directions after engineers detected severe cracking and ground collapse beneath a 194-year-old structure. It remains closed today, with no reopening date confirmed.

The 2026 RCC theme, The Highway Collapse, is not fiction. It’s the news. Your AI-controlled game is a response to a real emergency.

What is Path C? 🧠

Path C = Path B + AI Pose Control

  • Path B builds a rescue game using MRT Scratch AI Studio block coding
  • Path C does everything in Path B — research the theme, design a story, and build a Scratch rescue game — plus you train a Google Teachable Machine AI model
  • Your trained AI model connects to the game so that body movements control the main character or key sprites — instead of a keyboard or mouse
  • For example: raise both arms to jump, lean left to dodge debris, crouch to avoid danger
💡 The game complexity does not need to be higher than Path B. The key difference is how the game is controlled — by your trained AI, not by keys.

How to Participate — Step by Step

1
Same as Path B

Register

Sign up at roboticscodingchallenge.org/register-hub. Get instant access to the MRT eLearning platform with all tutorials included.

2
Same as Path B

Complete the eLearning Courses

Work through the Scratch coding courses on the MRT eLearning platform. Learn how to build interactive games with events, conditions, loops, and variables.

3
Same as Path B

Research the Theme & Design Your Game

Investigate the Highway Collapse scenario. Your research must shape your game — the story, characters, problem, and goal must connect to the real-world theme.

4
⭐ Path C Only — AI Step

Train Your Google Teachable Machine Model

Go to teachablemachine.withgoogle.com → choose Pose Project. Record yourself in at least 3–4 different body poses. Train the model. Export the model URL — you’ll need it for submission.

5
⭐ Path C Only — AI Step

Build the Game & Connect Your AI Model

Build your rescue mission game in MRT Scratch AI Studio. Import your Teachable Machine model and program the sprites so that pose inputs control the main character or key game actions — not the keyboard.

6
Same as Path B

Record a Video & Submit

Record a max 3-minute video showing your AI pose control working live in the game. Submit your video link, Teachable Machine model URL, and a short written description.


Your Mission 🎯

Build a rescue mission game based on the Highway Collapse theme. Research the real problem, design a story, and use MRT Scratch AI Studio to code it — then train a Google Teachable Machine Pose model and connect it so body movements control your game.

Your project must include:

  • A working rescue game built in MRT Scratch AI Studio (block coding)
  • A trained Google Teachable Machine Pose Classification model
  • At least the main character (or key sprites) controlled by AI poses — not keyboard
  • At least 3 distinct trained poses (4+ recommended)
  • A clear rescue goal connected to the Highway Collapse theme
  • Evidence of research that shaped the game design
🧠 AI is the primary input method — not an add-on. A student watching your video should be able to see you moving your body to control the game in real time.

Pose Control Ideas 🕺

Train these poses in Google Teachable Machine and map them to in-game actions:

⬆️

Both arms raised → Jump / Signal

Character jumps over collapsed debris or sends a rescue alert

↔️

Lean left / right → Move

Steer the character left or right to dodge falling rocks

Hand pushed forward → Stop

Halt the character before an unstable or dangerous area

🦆

Crouching → Duck

Avoid low-flying obstacles or ducking under collapsed sections

🙋

One arm raised → Request help

Call in a rescue team or trigger a special game event

🚀

Your own pose → Your own action

Design any pose that makes sense for your rescue game story


Allowed Tools 🛠️

MRT Scratch AI Studio — via eLearning Platform

A school-safe version of Scratch with Google Teachable Machine integration built in. Access via your eLearning account after registration.

🧠

Google Teachable Machine — Pose Project only

Free browser-based AI training tool. No account needed. Train at teachablemachine.withgoogle.com → select Pose Project.

Not accepted: Image Classification, Sound Classification, TensorFlow, PyTorch, Unity, Python, JavaScript, or Scratch (scratch.mit.edu). Pose Classification only.

How Your Project Is Scored 📊

Criteria 1 — Beginning 2 — Developing 3 — Accomplished 4 — Exemplary Score
AI Model Quality No model or unrecognisable 1–2 poses, inconsistent 3+ poses recognised accurately Stable model, diverse training, 4+ poses ×5 = 20
AI Integration AI not connected to game Partial connection, unstable AI reliably controls main character Seamless, creative AI-to-game integration ×5 = 20
Game Design No working game Basic movement, no clear goal Clear rescue goal, obstacles, scoring Polished game with levels and progression ×5 = 20
Research Connection Not related to theme Loosely referenced Research clearly shapes game design Deep, realistic rescue scenario ×5 = 20
Presentation No explanation Brief or unclear Clear demo of AI + game play Full walkthrough: AI training + live gameplay ×4 = 20

Total: 100 points


Submission Requirements 📦

  • A video link (YouTube or Google Drive) — max 3 minutes
  • Your Teachable Machine model URL (export from Teachable Machine)
  • A written description300+ characters
📹 Your video must show pose control working live — move your body on camera while the game responds. Voice explanation is strongly recommended.
📎 For upload tips → Submission Guide

Australian Curriculum Connections 🇦🇺

Digital Technologies
AI model training, data collection, algorithmic game control, input/output design
Design & Technologies
Iterative design, prototyping, testing, and evaluating
Health & PE
Body awareness, movement, physical engagement with technology
Critical & Creative Thinking
Designing AI-controlled solutions, evaluating model performance

Aligned with Australian Curriculum V9.0

Ready to train your AI and build a rescue game?

“Train your AI. Move your body. Save lives through code.”

Register to get instant access to all tutorials — Teachable Machine setup, MRT AI Studio guides, and project examples are all included.

Scroll to Top
Australian Government UNSW built Environment MRT Robotics Partnership Inquiries Your Logo Here
Organised In Partnership With UNSW Built Environment
Supported By
🔹 UNSW Built Environment Women in Construction Project 🔹 Department of the Prime Minister and Cabinet’s Office for Women 🔹 MRT Robotics (mrtrobotics.com) e-learning system

Corporate sponsors are acknowledged for supporting specific RCC events, awards, programs or student initiatives. Sponsor recognition does not imply endorsement by participating schools, public institutions or universities. Names and logos of institutions are used only where permission has been granted.