Abstract:Dynamical Systems (DS) are an effective and powerful means of shaping high-level policies for robotics control. They provide robust and reactive control while ensuring the stability of the driving vector field. The increasing complexity of real-world scenarios necessitates DS with a higher degree of non-linearity, along with the ability to adapt to potential changes in environmental conditions, such as obstacles. Current learning strategies for DSs often involve a trade-off, sacrificing either stability guarantees or offline computational efficiency in order to enhance the capabilities of the learned DS. Online local adaptation to environmental changes is either not taken into consideration or treated as a separate problem. In this paper, our objective is to introduce a method that enhances the complexity of the learned DS without compromising efficiency during training or stability guarantees. Furthermore, we aim to provide a unified approach for seamlessly integrating the initially learned DS's non-linearity with any local non-linearities that may arise due to changes in the environment. We propose a geometrical approach to learn asymptotically stable non-linear DS for robotics control. Each DS is modeled as a harmonic damped oscillator on a latent manifold. By learning the manifold's Euclidean embedded representation, our approach encodes the non-linearity of the DS within the curvature of the space. Having an explicit embedded representation of the manifold allows us to showcase obstacle avoidance by directly inducing local deformations of the space. We demonstrate the effectiveness of our methodology through two scenarios: first, the 2D learning of synthetic vector fields, and second, the learning of 3D robotic end-effector motions in real-world settings.

Modeling of Nonlinear Dynamical Systems Based on Deterministic Learning and Structural Stability

A Duplex Neurodynamic Learning Approach to Modeling Nonlinear Systems

Stability Analysis of Large-scale Nonlinear Systems with Hybrid Models

Deterministic Learning and Data-based Modeling and Control

A New Measure of Dynamic Similarity for Nonlinear Systems Based on Gap Metric and Deterministic Learning Theory

Data Based Identification and Prediction of Nonlinear and Complex Dynamical Systems

Regression-based Identification and Order Reduction Method for Nonlinear Dynamic Structural Models

A Modeling Method for Process Systems and Studies on the Dynamical Behavior

Learning Stable Deep Dynamics Models for Partially Observed or Delayed Dynamical Systems

Deterministic Learning and Nonlinear Observer Design

Analyzing Of Hopf Bifurcation Based On Deterministic Learning

Time Series Classification of Dynamical Systems Using Deterministic Learning

Deterministic learning from sampling data.

Learning Networked Dynamical System Models with Weak Form and Graph Neural Networks

Learning Control-Oriented Dynamical Structure from Data

Learning Linearized Models from Nonlinear Systems with Finite Data

New Results on Deterministic Learning of Sampled-data Nonlinear Systems

A Data-Driven Framework for Learning Hybrid Dynamical Systems.

Learning Dynamical Systems Encoding Non-Linearity within Space Curvature

Data-driven Effective Modeling of Multiscale Stochastic Dynamical Systems

Fault diagnosis for a class of nonlinear uncertain systems using deterministic learning approach