To install the required libraries run <pip install -r requirements.txt> Once the required libraries are installed, to run the full pipeline use
To use JAX with the GPU, please follow instructions from the official page linked below to install the right version of JAX. https://jax.readthedocs.io/en/latest/installation.html
The codes used in the paper for unicycle model experiments are found in the Simulations folder. It is organized as shown below.
Simulations/
- data/ - Data for the unicycle dynamical system
- mlp_jax.py - An implementation of the multilayer perceptron network using JAX libraries
- model_learning.py - An implementation of value function learning pipeline using JAX libraries
- helper_functions.py - Useful functions for computing tracking cost and input for the unicycle system
- generate_data.py - Contains the ILQR modules and unicycle simulation function and other helper functions
- run_exp.py - Example usage for running a full scale experiment on a simple unicycle system
- Testing_mlp.ipynb - Contains an example for generating trajectories for the unicycle system and visualizing them
To generate data, please take a look at how to call the functions as shown in the Testing_mlp.ipynb notebook. The generated data is saved to the data/ folder.