Multi-Agent Coordination Project

Project Goals

The learning goal for this project is to get you to achieve the following:

• Understand the system workflow for the multi-agent pickup and delivery (MAPD) problem
• Build confidence in controlling the Crazyflies using different control interfaces (e.g. position, velocity setpoints)
• Improve coding skills in Python
• Explore the multi-agent path finding (MAPF) problem in the context of the Hummingbird project

Project Description

• Given a defined task list (including pick and drop locations for N deliveries), operate M Crazyflie agents to complete the defined task list as quickly as possible.
• Here, number of tasks (N) is 4, and number of Crazyflie agents (M) is 3. See images below for a graphical representation of the task list (specific information is provided in the config.yaml file)
• To make the project more tractable in the available time:
  • We will complete the task list in the flight area WITHOUT physical obstacles (i.e., bin racks)! However, you should plan your solution as if they were present.

---

Setting up your PC

📌 Of course, you should have already set up your PC by now (if so, just skip to **Step 3**), but just in case you need a refresher 🙂

You can complete this assignment on any computer OS: Mac, Windows, Linux, etc. All you need is Python 3.6+.

Step 0 (Optional)

Install Visual Studio Code (I strongly recommend using this, if you don’t already do. It’s the best IDE in my humble opinion)

• Follow this to install it: Download Visual Studio Code - Mac, Linux, Windows

Step 1

Get Python 3 installed on your computer (if you do not already have it)

• To check if you have it, open your terminal (or CMD, PowerShell, etc.) and type:

  C:\Users\kmbanisi> python --version    # <--- type this
  Python 3.8.3                           # <--- you should see something like this

• If you don’t have Python installed, follow this tutorial to install it: Python 3 Installation & Setup Guide - Real Python

Step 2

Create a virtual environment for this assignment [This is technically optional, but I strongly recommend it]

• Follow this tutorial (Windows): Python Virtual Environments: A Primer - Real Python
• Follow this tutorial (Linux): Create a Python virtual environment in the Linux terminal

Step 3

Get the assignment repository from Github

• Clone this repository: https://github.com/Olin-HAIR-Lab/multi-agent-coordination-project.git

• Follow this instruction to clone:

  # first: cd to your desired folder
  
  # then: git clone the project
  git clone https://github.com/Olin-HAIR-Lab/multi-agent-coordination-project.git
  cd multi-agent-coordination-project

Step 4

Install all required Python packages

# first: make sure you have activated the virtual environment (if you used one). See step 2 tutorial

# cd to the project folder
cd multi-agent-coordination-project

# install all required packages from requirements.txt
python -m pip install -r requirements.txt

Step 5

There are TWO ways to interact with this project:

1. You can work out of the Jupyter Notebook directly from VSCode, in that case, you’ll use the main.ipynb file [This is what I recommend]
2. You can work out of a normal python script. In this case, run the following:

python main.py

What to do and what to submit

• Task 1: Review and understand Task and Agent definitions in STEP 1 in the main.ipynb file

  • N.B.: There’s nothing to implement here. Just understand what we’re doing.

• Task 2: Review and understand the multi-agent task assignment implementation in STEP 2 in the main.ipynb file

  • N.B.: There’s nothing to implement here. Just understand what we’re doing.

• Task 3: Run the existing code:

  • Task 3.1: Run it in simulation and obtain plots such as below:

    • Solution Plots:

      

      

  • Task 3.2: Run it on the Crazyflies (make sure to set the right parameters in the config.yaml file.

    • Obtain plots as shown above and also record a video

• Task 4: Implement a (simple) multi-agent path finding algorithm (STEP 3 in the main.ipynb file) that computes a conflict-free paths/trajectory for each agent.

• Task 5: Integrate your (existing) trajectory-following algorithm from the Module 2 project (STEP 4 in the main.ipynb file)