Preliminaries

Table of contents

1. Rules of the game
   1. Warm-up exercises, not graded
   2. Weekly individual exercises, graded
   3. Finals exercises, graded
   4. Extra exercises from previous years, not graded
2. Plagiarism and dishonest conduct
3. AI usage
4. Getting started
   1. Python
      1. Resources
   2. Docker
      1. Resources
      2. Checklist
   3. Git & GitHub
      1. Creating a GitHub account
      2. Setting up Git
      3. Resources
      4. Checklist
   4. VS Code
5. Support

---

Rules of the game

The programming exercises are a great way to understand and apply what you learn in class.

There are four types of exercises:

• Warm-up exercise(s), not graded;
• Weekly individual exercises, graded (will account for 30% of the final grade);
• Finals group exercises, graded (will account for 70% (30 + 40) of the final grade);
• Extra exercises from previous years, not graded.

Warm-up exercises, not graded

Just ensuring everything is working as expected.

Weekly individual exercises, graded

The weekly exercises must be solved individually. They will involve relatively short programming tasks related to the topics covered in class. You will submit your solutions via your private GitHub repository. Detailed submission instructions are provided later.

Important information:

• Each exercise has a limited submission window for 3-4 weeks.
• Only the last VALID submission (i.e., processed by the server regardless of the outcome) within the exercise time window counts toward your exercise grade. A missing submission counts as 0.
• Your score will be determined by the performance metrics specified in the exercise description. For students who wish to pursue higher performance(e.g. in solving time), we provide reference values for the full grade based on the TA solution.
• Only your best N-1 exercise results contribute to 30% of your final grade, where N is the total number of graded weekly exercises (N=5).
• For each exercise, you have a limited number of server submissions. You can test locally as much as you want. Consider this a learning experience for robotics; field tests are expensive.
• Here you can see a temporary schedule (will be completed on the fly):

Exercise ID	Topic	Evaluation opens	Evaluation closes	Deadline status	Available Submissions
02	Graph Search	23rd of September	20th of October (23:59 CET)	confirmed	10
03	Informed Graph Search	23rd of September	20th of October (23:59 CET)	confirmed	10
04	Dynamic Programming	30th of September	27th of October (23:59 CET)	tentative	10
05	Steering (Dubins)	14th of October	3rd of November (23:59 CET)	tentative	10
06	Collision Checking	21th of October	10th of November (23:59 CET)	tentative	10

• Most of these exercises are already accessible before the official evaluation opens. You are free to engage with them earlier, but be aware that changes may occur up until the official opening. After the evaluation opens, the exercise is “frozen”, and you can solve and submit it.

Finals exercises, graded

These are solved in small groups of up to 3 people, with the same modality as the weekly exercises. They will be a bit more involved content-wise. Instructions on group formation are provided via Piazza.

Exercise ID	Topic	Evaluation opens	Evaluation closes	Deadline status	Available Submissions
13	Satellite Docking	11th of November	1st of December (23:59 CET)	tentative	10
14	Multi-agent Goal Collection	2nd of December	22nd of December (23:59 CET)	tentative	10

Extra exercises from previous years, not graded

You can have a look at the finals from last year and challenge yourself.

Exercise ID	Topic	Evaluation opens	Evaluation closes
07	Optimization	-	-
08	Driving Games	-	-
09	PDM4ARocket Explorer	-	-
10	Robot Runners	-	-
11	Spaceship	-	-
12	Highway Driving	-	-

Plagiarism and dishonest conduct

We take plagiarism and dishonest conduct very seriously. The exercises are meant to be solved individually. We will check for plagiarism with automated software and human help. Violating the ETH plagiarism etiquette will result in disciplinary actions as per ETH regulations.

“C. The consequences of plagiarism Pursuant to Art. 2 Para. b of the ETH Zurich Disciplinary Code (RSETHZ 361.1) plagiarism constitutes a disciplinary violation and will result in disciplinary procedures. Detailed information regarding these procedures and their jurisdiction may be found in the ETH Zurich Disciplinary Code (RSETHZ 361.1 / www.rechtssammlung.ethz.ch).”

AI usage

AI coding assistants are becoming a legitimate part of a modern robotics workflow. Therefore we encourage you to learn how to use them to accelerate your development.

The following assistants are preinstalled in the development environment:

• VS Code extensions: Claude Code, Codex, and GitHub Copilot;
• CLI tool: Antigravity (agy, backed by Google Gemini).

Each of these requires your own account, and we unfortunately cannot provide any paid quota. We recommend starting with the free plans and only then deciding whether a subscription is worthwhile for you. You are also free to use any other assistant.

What is acceptable. Using an assistant for the supporting parts of the work: looking up documentation and APIs, writing tests, debugging, and profiling or optimizing an implementation you have already designed.

What is not acceptable. Delegating the design and implementation of an algorithm to an AI assistant. The purpose of these exercises is that you can reason about, implement, and debug planning algorithms.

We will check for excessive reliance on AI and for a lack of familiarity with the content of the exercises.

Getting started

We will use:

• Python as programming language;
• Docker as environment containerization (but you won’t see it);
• Git as version control system;
• GitHub as code hosting and assignment distribution and submission platform;
• VS Code as code editor.

If these tools are new to you, do not panic. We require only very basic use of most of them. It’s a good time to start learning these tools, as they are widely adopted in modern robotics.

If you get stuck, pair up with experienced colleagues. If that doesn’t solve the problem, reach out to the instructors on Piazza or during office hours.

If you are using a Linux-based OS, the process should be straightforward. Windows and macOS may have more hiccups, but they are supported as well.

Python

Python will be the programming language adopted in this course. You don’t need to install Python on your local machine because you’ll work in a Docker container with Python preinstalled (more information below).

Resources

If you want to learn Python basics, here are some resources to get you started:

• Python Tutorial, in particular make sure to go through the basic data structures (tuples, lists, dictionaries, sets,…), loops (while, for,…), conditional statements (if-else), functions, classes and objects.
• Official documentation

Docker

We will run the exercises in a containerized environment, a technique widely used in modern software development. Read the Resources section for more information on Docker.

Now let’s install it on your computer:

If you are using Linux, you can choose to either install Docker Engine or Docker Desktop, but NOT both. Docker Engine is the core container runtime and is bundled with the Docker CLI docker, while Docker Desktop is Docker Engine + GUI + extra tools.

👉 For Linux users of this course, Docker Engine is recommended for its simplicity and lower resource usage.

• Install Docker Engine on Linux, or
• Install Docker Desktop on Linux

If you are using Mac, follow the instructions below to install Docker Desktop:

• Install Docker Desktop on Mac

If you are using Windows, follow this YouTube video to install Docker Desktop:

• Install Docker Desktop on Windows

Resources

If you want to learn more about Docker, check out the following resources:

• Docker intro

Checklist

• I have installed Docker on my computer
• I can run without errors the Docker hello-world (docker run hello-world)

Git & GitHub

Git is a version control software. GitHub is a cloud-based platform that hosts Git repositories. Please see the Resources section if you are new to them. You will need Git on your computer and a GitHub account.

Creating a GitHub account

If you do not already have a GitHub account create a new one here.

Setting up Git

For Linux or Mac users:

• Download and install Git for your OS from this site.
• Set up Git and authenticate with GitHub following the instructions here. You are free to pick the connecting and credential caching options that you are most comfortable with.

For Windows users,

• Download and install Git for Windows from here, which comes with the Git Credential Manager (GCM).
• The simplist way to connect to GitHub on Windows is over HTTPS and use GCM to store your Git credentials. See here for detailed instructions.

Notes for Windows users:

• VS Code with Dev Container extension(see below) automatically forwards your cached credential from Windows to the container. This means you can perform all Git operations both on Windows and inside the container without re-authenticating.
• If you clone the repository on Windows, remember that our Linux-based container will access the files through the Windows filesystem. This causes significant overhead for file-intensive operations (e.g., I/O, dependency installation). For this course, the impact is negligible since no bulky I/O operations are expected.
• Cloning the repository inside WSL instead of Windows avoids the overhead for file-intensive operations. However, it requires configuring GCM on WSL separately. If you are interested in following this approach (not necessary for this course). instructions are available here.

Resources

• Git Guide
• GitHub tutorial

Checklist

• I can run git version on my terminal without errors.
• I have a GitHub account.
• I know how to use basic git commands git clone, git add, git commit and git push.

VS Code

Using an IDE is not necessary. But it provides a good set of tools that speeds up development (code navigation, debugging,…). Moreover, we will provide environment configurations that will make your life easier.

We recommend using VS Code. You can find installation instructions here. After installing VS Code, also install the Dev Containers extension from here.

There are many other good IDEs for Python (PyCharm, Atom, Spyder,…) that should work just fine if you know how to replicate the development environment in .devcontainer/ exactly. We won’t support them officially.

Support

Use the forums on Piazza for general questions: this way, you can help other students who experience the same issues.