Guaranteed Robust Nonlinear MPC via Disturbance Feedback

This repository contains scripts for running robust control simulations and experiments on pendulum, quadrotor, and rocket systems.

---

Closed-loop rocket simulation

Prerequisites

• Python 3.11+
• Required packages: see requirements.txt
• Some scripts may require additional solvers (e.g., IPOPT)

Setup with virtual environment

# Create a virtual environment (in folder .venv)
python -m venv .venv
# Activate the virtual environment
# On Linux/macOS:
source .venv/bin/activate
# On Windows:
.venv\Scripts\activate

# Install dependencies:
pip install -r requirements.txt

Solver Generation

Run the script corresponding to your system to generate the necessary solvers:

python solver/generate_osqp_rockETH.py

Then, install:

cd build/osqp_fast && python -m pip install --use-pep517 . && cd ../..

Warning: Always make sure to regenerate the solvers if you run new experiments. In particular, no error will be thrown if you use a solver on the wrong problem.

Running Experiments

Most of the scripts starting with main_ are used to run experiments. Options can be specified via command-line arguments. For example:

python main_rocket_robust_closed_loop.py --run

For instance, you can plot the results of the last experiment using:

python main_rocket_robust_closed_loop.py

Notes

• All scripts save results in their respective output folders. The corresponding plot function will retrieve the last saved data and plot it.