This repository is a catch-all simulation environment for studying controllers (which may have bandwidth limitations) for dynamical systems.
dyn_sim
- core codectrl
- implemented controllers with subfolders corresponding to each system. General control code lives in the parent folder (e.g.mpc
).quad
- Inner/Outer Loop PD Controller
- [REFACTORING] Safe Multirate Controller
- [WIP]
segway
sim
- simulator codedyn_sys
- implemented dynamical systemsplanar
- systems that are visualized in the plane- [WIP] segway
spatial
- systems that are visualized in Euclidean space- quad
util
- common utility functions across codebase
scripts
- scripts for running simulations
This repo uses the optimization package Gurobi, which is a commercial product (but is available to academics for free). This section details (hopefully) headache-free installation of Gurobi for academics. If you are not an academic, wait for alternate free optimizer support (e.g. scipy
).
First, register using your academic credentials on the website (instructions for the academic license are linked here).
Second, we will only require the package gurobipy
, so we will install that with the requirements file in the next section. However, we must license our installation, so install the license tools package here. Extract the executable file grbgetkey
into any directory.
Third, on the Gurobi website, you should be able to view your current licenses. If you click on your License ID, you will come to a page with Installation instructions, which should include a line that looks like:
grbgetkey xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
Copy this command, and in your terminal, run the command
<full_parent_path>/grbgetkey xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
with the correct parent path and license key. Save the license file somewhere logical, such as the home directory. gurobipy
should be usable on your machine after its installation via pip
.
This repo is designed to be run in a conda
virtual environment with dependencies managed by pip
and versions frozen in a requirements file (versions last frozen May 23, 2022).
To make an environment, make sure you have conda installed and then run the command:
conda create --name <env_name> python=3.10.4
Anytime after initializing the environment, activate before using project files:
conda activate <env_name>
To install the dependencies in the project, after activating the environment for the first time, run
pip install -r requirements.txt
For people developing on the environment, style-checking and static type-checking is performed using pre-commit hooks. Make sure you have pre-commit
installed (separate command from dependency installation):
pip install pre-commit
pre-commit install