🚕AI +(Python)

Description

This flexible activity is designed to teach instructors how to harness large language models (LLMs) like ChatGPT to generate and refine JavaScript code that drives a four-motor robot. Each self-contained activity focuses on a core AI or robotics concept, pairing a clear learning objective with hands-on prompt exercises and real-world testing. Instructors can pick and choose which activities to run and in what order to fit their schedule and learners’ needs.

Overview

This guide covers:

• Generating PWM commands for the ROV.

• Understanding how PWM settings affect motor drive and direction.

• Running user-generated Python using LLM output based on guided input.

In order to run python, you will need to have an IDE and python compiler. If desired, you can use Visual Studio Code Code on your computer.

Follow this link https://code.visualstudio.com/docs/setup/setup-overview and use the steps related to your operating system.

Once you have the IDE installed, follow these instructions to install python, https://code.visualstudio.com/docs/python/python-tutorial.

Bleak library

• This library must be installed first https://github.com/hbldh/bleak

• From the command prompt run: Run “pip install bleak”

RoverLLM.py

The activities in this lesson are based on the contents of this file. When prompted below, you can use the copy feature to save a copy of the file to your device

import asyncio
import gc
from bleak import BleakClient, BleakScanner

DEVICE_NAME = "ROV_7446B3EA794D"
MAILBOX_CHARACTERISTIC = "0000ffd1-0000-1000-8000-00805f9b34fb"
COMMAND_CHARACTERISTIC = "0000ffd2-0000-1000-8000-00805f9b34fb"

client = None  # Global client variable

async def find_device_by_exact_name(name):
    print("Scanning for BLE devices...")
    devices = await BleakScanner.discover(timeout=10.0)
    for d in devices:
        if d.name == name:
            print(f"Found device: {d.name} ({d.address})")
            return d.address
    print(f"No device found with name: {name}")
    return None
    
async def initialize():
    await MoveMotors(-1, -1, -1, -1)
    await MoveMotors(0, 0, 0, 0)
    
async def MoveMotors(m1, m2, m3, m4):
    motors = [max(-100, min(100, val)) & 0xFF for val in (m1, m2, m3, m4)]
    payload_mailbox = bytearray(motors)
    payload_command = bytearray([0x13, 0x00])

    await client.write_gatt_char(MAILBOX_CHARACTERISTIC, payload_mailbox, response=False)
    await client.write_gatt_char(COMMAND_CHARACTERISTIC, payload_command, response=False)
    print(f"Motors moved: [{m1} {m2} {m3} {m4}]")

async def StartProgram():
    print("Move Forward at 50% for 2 seconds")
    await MoveMotors(50, 50, 50, 50)
    await asyncio.sleep(2)
    print("Stop")
    await MoveMotors(0, 0, 0, 0)
    await asyncio.sleep(2)

async def main():
    global client
    address = await find_device_by_exact_name(DEVICE_NAME)
    if not address:
        return

    client = BleakClient(address)
    try:
        await client.connect()
        if not client.is_connected:
            print("Failed to connect.")
            return
        print(f"Connected to {DEVICE_NAME}")

        await initialize()
        await StartProgram()

    except Exception as e:
        print(f"Error: {e}")
    finally:
        try:
            print(f"Attempting disconnect... connected: {client.is_connected}")
            await client.disconnect()
            await asyncio.sleep(1)
            gc.collect()
            print("Disconnected and cleaned up.")
        except Exception as e:
            print(f"Error during disconnect: {e}")

if __name__ == "__main__":
    asyncio.run(main())

Activity 1 – LLM Setup for Coding

Objective: Configure the LLM to output code using your ROV-specific API.

Prerequisites:

1. Open VS Code and make a new Python file called Activity1.py. based on RoverLLM.py

2. Copy RoverLLM.py to activity1.py

3. Change the DeviceName (line 5) to match the Device Name of your device.

Prompt to ChatGPT:

I have a robot with two equal-but-opposite motors driving its wheels. 
It exposes two Python functions
•	a local function MoveMotors(M1, M2, M3, M4) where each motor speed M# is –100…100 
•	asyncio.sleep(x) which delays the program for x seconds
•	M1 and M2 are on the same side and should never have opposing motion
•	M3 and M4 are on the same side and should never have opposing motion


Examples:
MoveMotors(100, 100, 100, 100); // forward at full speed for 1s  
MoveMotors(-100, -100, -100, -100); // backward at full speed for 1s
MoveMotors(100, 100, -100, -100); // turns left   
MoveMotors(-100, -100, 100, 100); // turns right   
MoveMotors(0, 0, 0, 0); // stops all motors   

I want you to help me generate Python code from prompts to perform robot activities.
•	Use simple, non-coder-friendly constructs.
•	Provide clear comments and explain each part of the code you generate.
•	You should implement a async def StartProgram() function that contains your logic. 
•	You should only return the StartProgram function definition, no other code or imports are needed.
•	All MoveMotor function calls should be asynchronous
•	before each command, add a print to write the action to the console
•	start with an example that move forward for 5 seconds at 30% and stops.

This prompt should return back some sample code you can try.

Next Steps

1. Replace the StartProgram function with the output from ChatGPT

2. Execute the program in VS Code by click the run python file from the menu bar.

Activity 2 - Sensing & Reactive Behavior

Objective: User Distance to control movement

Prerequisites:

• Open VS Code and make a new Python file called Activity2.py. based on RoverLLM.py

• Copy RoverLLM.py to activity2.py

• Change the DeviceName (line 5) to match the Device Name of your device.

Prompt:

Write code that:
1.	Move forward at 30%
2.	Backup at 30 percent for 1 second
3.	Turn right at 30 percent for 1 second
4.	Continue the cycle 

Next Steps

1. Replace the StartProgram function with the output from ChatGPT

2. Execute the program in VS Code by click the run python file from the menu bar.

Activity 3 - Random Movements

Objective: generate an algorithm with ordered steps based on user input.

Prerequisites:

• Open VS Code and make a new Python file called Activity3.py. based on RoverLLM.py

• Copy RoverLLM.py to activity3.py

• Change the DeviceName (line 5) to match the Device Name of your device.

Prompt:

Write code that:
1.	Move forward at 30% for 3 seconds.
2.	Stop for 2 seconds
3.	Back up for 3 seconds
4.	Stop for 1 second 
5.	Turn right for .5 seconds
6.	Stop 
7.	Turn left for .5 seconds
8.	Stop
9.	Spin in a circle for 5 seconds at 30%
10.	Spin in the opposite direction for 5 seconds at 30%
11.	Stop 

Next Steps

1. Replace the StartProgram function with the output from ChatGPT

2. Execute the program in VS Code by click the run python file from the menu bar.

Activity 4 – Move in a Square

Objective: Generate a algorithm to move the robot in a square

Prerequisites:

• Open VS Code and make a new Python file called Activity4.py. based on RoverLLM.py

• Copy RoverLLM.py to activity4.py

• Change the DeviceName (line 5) to match the Device Name of your device.

Prompt:

Write code to:
Drive in a square, moving forward at 30% for 2 seconds, then turn for 4 seconds at 30%.

Next Steps

1. Replace the StartProgram function with the output from ChatGPT

2. Execute the program in VS Code by click the run python file from the menu bar.