Thymio Robotics Project

A course project for MICRO-452 as part of MA-RO1 at EPFL, December 2022.

Designed to run on the Thymio robot. Written in Python, navigates the robot through a dynamic obstacle scene.

Features

• 📷 Realtime image processing for object detection using OpenCV
• 🤖 A Big Brain to make complex decision
• 🗺️ Optimal pathfinding using Dijkstra's algorithm
• 🚗 PID based motion control
• 📱 Realtime WebSocket based control server & Web UI
• 🧵 Asynchronous event loop
• 🚨 Cool lights
• 🔊 Cool sounds

Prerequisites

In order to run the project, you wil need:

• A Thymio robot (or the simulator)
• A USB camera
• Python 3.10 or higher

Optionally, to use the real-time visualisation tool, you will also need a modern version of Node.js® (>=16).

Creating a virtual environment

It's recommended to create a "venv" to isolate the Python runtime and dependencies from your system's global installation for more reliable results.

$ python -m venv venv

To activate the virtual environment, run the script corresponding to your terminal (such as .ps1 for PowerShell):

$ .\venv\scripts\Activate.ps1

This can be deactivated by simply running:

$ deactivate

Installing dependencies

While the virtual environment is activated, run the following command to install all the required dependencies:

$ pip install -r requirements.txt

Setting up the visualisation tool

The visualisation tool (referred to here as the Web UI) is a Node.js application that can be used to display the robot's camera feed. detected objects and a map with path-finding information in real-time. For best results, it is recommended to build the application prior to running it.

$ cd ui
$ npm install
$ npm run build

If there are any issues, try installing dependencies using the pinned versions from the lock file.

$ npm ci

Figure: Screenshot of the web interface

Usage

The application should be run from the root directory of the project.

$ python -m app

This will attempt to connect to the Thymio driver, lock a Thymio node and start the main event loop. The control server is run on port 8080.

First enter whether you would like to connect to a second Thymio node (Y/n), then select the corners of the map and the two landmarks in the GUI window that appears (instructions will also be printed in the terminal).

Using the Web UI, you can select a destination point and the robot will navigate autonomously to that point, avoiding expected or unexpected obstacles.

To exit the application, press Ctrl+C in the terminal or use the stop button in the Web UI.

Starting the visualisation tool

In a separate terminal, cd back into the ui directory and run the following:

$ cd ui
$ npm run preview

This will start a local web server on port 4173. You can then open the application in your browser by navigating to http://localhost:4173.

The connection configuration should bet set to localhost:8080 by default. This assumes that the control server is running locally.

The control server supports any number of clients, so you can open the Web UI in multiple tabs or windows to see the same information. Each client receives an initial snapshot of the application State object, and then receives incremental updates (or patches) as the state changes.

Starting the notebook

The notebook report is used to explain the project and present the results. It can be run using jupyter-lab.

$ pip install jupyterlab
$ jupyter-lab

Configuration

All application configuration is stored in the config.py file. To use a simulated image, for example, set USE_LIVE_CAMERA to False.

Authors

• Marcus Cemes
• Pablo Palle
• Adrien Pannatier
• Carolina Rodrigues Fidalgo

License

This project is released under the MIT license.

See LICENSE.