Skip to content

DecodEPFL/outputSLS

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Boosting the performance of nonlinear systems through NN-based output-feedback controllers

PyTorch implementation of our NN-based control framework presented in the
paper titled "Parametrizations of All Stable Closed-Loop Responses: from Theory to Neural Network Control Design"

Installation

git clone https://github.com/DecodEPFL/outputSLS.git

cd outputSLS

python setup.py install

Examples:

The environment consists of two robots that need to achieve some predefined task while avoiding collisions between them and the obstacles. In the first case, the robots need to pass through a narrow corridor, while in the second case they have to visit different goals in a predefined order.

Mountains

The following gifs show the trajectories of the 2 robots before training (left), and after the training of the proposed controller (right). The agents need to coordinate in order to pass through a narrow passage while avoiding collisions between them. The initial conditions used as training data are marked with ○, consisting of s = 100 samples. The shown trajectories start from a random initial position sampled from the test data and marked with ○.

robot_trajectories_before_training robot_trajectories_after_training

Waypoint-tracking

The following gifs show the trajectories of the 2 robots before training (left), and after the training of the proposed controller (right). The agents need to visit different waypoints in a specific order.

robot_trajectories_before_training robot_trajectories_after_training

License

This work is licensed under a Creative Commons Attribution 4.0 International License.

CC BY 4.0

References

[1] Clara Galimberti, Luca Furieri, Giancarlo Ferrari-Trecate. "Parametrizations of All Stable Closed-Loop Responses: from Theory to Neural Network Control Design," 2024.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages