In this project on Github I coded up some tutorial concepts in control theory like observability, controllability, and the linear quadratic regulator using the example environment of a linear pendulum fixed to a cart. I also made fun Jupyter notebook movies to visualize the results.
Also included under this project are the slides I did for a short class project covering reinforcement learning (RL) for CartPole from the OpenAI lab. It’s nice to contextualize model-free RL methods for control within a familiar environment where we have covered the control theory basics.
TODO: Turn this into a Google Colab https://colab.research.google.com/
Embedding the notebook as a static webpage means there’s not a lot of fun to have with the movies so come back soon for a dynamic version of this post.
