Home > Using RLAI with Raspberry Pi

Using RLAI with Raspberry Pi

Most of the RLAI functionality can run on a Raspberry Pi. This opens possibilities for building physical systems with the Pi’s GPIO ports and then learning control policies for these systems with RLAI. See here for instructions on configuring an operating system on the Raspberry Pi. This page provides details on installing RLAI on the Pi for use within a Python project running the physical system (e.g., the cart-pole system described here).

Installing RLAI

Log in to the Pi via the desktop or SSH connection.
Install dependencies:
```
sudo apt install build-essential swig
```

Configure a virtual environment within the project:

cd /path/to/rpi/project
python3 -m venv venv
. venv/bin/activate
pip install -U pip

Clone RLAI and install it within the project’s virtual environment:
```
cd ~
git clone https://github.com/MatthewGerber/rlai.git
pip install rlai
```
This step will take quite a while, since pip will need to compile several of the RLAI dependencies from source.

Limitations

RLAI depends on Qt6 for graphical rendering of certain simulations and plots. These renderings will not be possible on the Pi.