<aside> <img src="/icons/layers_pink.svg" alt="/icons/layers_pink.svg" width="40px" /> This project aims to explore the role that AI can play in coding robotics in two parts…

Program for AI Robotics Engineering (PARE):

Focuses on the core capabilities of AI-programmed robotics, such as autonomous decision making, direct hardware control, iterative coding, and independent verification of success.

Scaffolded Program for AI Robotics Engineering (SPARE):

Builds on the foundation of PARE but introduces a network of manager and worker AI Assistants. This hierarchical system experiments with the execution of more complex tasks, that address multiple microcontrollers, by delegating them to sub-workers and abstracting them from the user.

</aside>

<aside> <img src="/icons/info-alternate_purple.svg" alt="/icons/info-alternate_purple.svg" width="40px" /> Details

• Authors: Jesse Gilbert and Liam Campbell
• GitHubs: https://github.com/theLiamFC/PARE & https://github.com/theLiamFC/SPARE
• Research guided by: Professor Chris Rogers and Professor Ethan Danahy </aside>

Program for AI Robotics Engineering (PARE)

<aside> <img src="/icons/map-pin_orange.svg" alt="/icons/map-pin_orange.svg" width="40px" /> This project explores the limits of AI capabilities in robotics engineering with the goals of…

• Minimizing the context specific training & knowledge given to the AI model. This enables the future possibility of “plug and play” work where our system could adapt to code new and unseen hardware.
• Reversing the traditional AI chatbot to user role with “human in the loop” programming. We want to test what an LLM is capable of when it’s given the tools to operate autonomously.
• Enabling AI autonomy to influence real world hardware. This allows for efficient testing of code and gives the AI opportunities to run independent tests without user input. </aside>

<aside> <img src="/icons/school_green.svg" alt="/icons/school_green.svg" width="40px" /> Educational Applications Could a student learn programming and problem-solving techniques by watching how an AI interacts with a robot? Our testing with P.A.R.E. indicates that an LLM in our framework has the ability to iterate its own designs, debug hardware, and even create unit tests for soft and hardware. We believe these are all essential engineering skills for work with robotics that could be learned by engaged students.

</aside>

Conceptual Overview

• ← Click here to read the custom instructions we have given to the OpenAI Assistant.
• ← Click here to get a closer look at how each of AI Alchemist’s four features are formatted through the OpenAI Beta Assistant function calling feature.

What can PARE do with a Lego SPIKE Prime?

Run SPIKE motors:

Hey SPIKE: Drive the robot forward.

This video showcases one of the most basic features of the AI Alchemist program which is its ability to directly run code on the LEGO SPIKE Prime.

This video showcases one of the most basic features of the AI Alchemist program which is its ability to directly run code on the LEGO SPIKE Prime.

Create Sensor Dependent Physical Outputs:

Hey SPIKE: Make a musical instrument that I can play.

This video is an example of the AI Alchemist creating an interactive musical instrument with the LEGO SPIKE Prime.

This video is an example of the AI Alchemist creating an interactive musical instrument with the LEGO SPIKE Prime.

Hey SPIKE: Follow the blue line on the table.

This video is an example of the AI Alchemist tackling a semi difficult programming task involving real life sensor based motor movement.

This video is an example of the AI Alchemist tackling a semi difficult programming task involving real life sensor based motor movement.