ASACK : Adaptive Safe Active Continual Koopman Learning for Uncertain Systems with Contractive Guarantees

TL;DR

This paper introduces ASACK, a framework combining adaptive Koopman learning, active data collection, and safety guarantees for real-time control of uncertain nonlinear systems.

Contribution

It proposes a unified, safe, and efficient online Koopman model refinement method with theoretical guarantees, suitable for real-time robotic applications.

Findings

The method achieves superior control performance in simulations and experiments.

It provides formal safety guarantees via robust MPC incorporating model error bounds.

The approach effectively handles distributional shifts and model uncertainties.

Abstract

Koopman operator theory provides a powerful framework for representing nonlinear dynamics through a linear operator acting on lifted observables, enabling the use of linear control techniques for nonlinear systems. However, Koopman models are typically learned from data and often degrade in performance under model uncertainty and distributional shifts between training and deployment. Although several works have explored online adaptation to address this issue, many rely on neural network-based updates that introduce significant computational overhead and lack formal safety guarantees, limiting their suitability for real-time and safety-critical robotic applications. In this work, we propose a unified framework for continual adaptive Koopman learning that enables safe and efficient online refinement of learned models during task execution. An autoencoder-based Koopman model is first learned offline and subsequently refined online through a contractive adaptation law, which provides theoretical convergence guarantees under distributional shifts and model uncertainty. To improve data efficiency and accelerate model refinement, the adaptation mechanism is integrated with an active learning strategy that drives the system to collect informative data while accomplishing task objectives. The resulting control problem is formulated as a nonconvex optimization problem incorporating both active learning objectives and safety constraints. We further derive theoretical bounds on model approximation error and show how these bounds can be incorporated within a robust Model Predictive Control (MPC) framework to provide formal safety guarantees. The proposed approach unifies learning, excitation, and safety within a single control framework without sacrificing real-time feasibility. Extensive simulation and experimental studies demonstrate superior performance compared to state-of-the-art baselines.