Task-Oriented Koopman-Based Control with Contrastive Encoder

TL;DR

This paper introduces a novel task-oriented Koopman control method that combines reinforcement learning and contrastive encoding to handle high-dimensional nonlinear systems, including pixel-based and real robot tasks.

Contribution

It extends Koopman control to high-dimensional, complex nonlinear systems using an end-to-end reinforcement learning approach with contrastive encoding, reducing model reliance.

Findings

Successfully applied to pixel-based tasks

Effective on real robot with lidar data

Extends Koopman control to complex systems

Abstract

We present task-oriented Koopman-based control that utilizes end-to-end reinforcement learning and contrastive encoder to simultaneously learn the Koopman latent embedding, operator, and associated linear controller within an iterative loop. By prioritizing the task cost as the main objective for controller learning, we reduce the reliance of controller design on a well-identified model, which, for the first time to the best of our knowledge, extends Koopman control from low to high-dimensional, complex nonlinear systems, including pixel-based tasks and a real robot with lidar observations. Code and videos are available \href{https://sites.google.com/view/kpmlilatsupp/}{here}.