Discovering Governing Equations of Robots from Data

Abstract

This paper introduces a novel method for discovering the governing equations of robots from data, employing Wide-Array of Nonlinear Dynamics Approximation (WyNDA). WyNDA utilizes an innovative algorithm rooted in adaptive observer techniques, facilitating the identification of both the structure and parameters inherent in the governing equations. To assess the efficacy of this approach, extensive simulations are conducted using a single-link joint robot system. The results showcase that the proposed WyNDA method achieves an impressive accuracy exceeding 97%, underscoring its effectiveness in precisely characterizing the dynamics of robotic systems. This research contributes to the advancement of techniques for extracting governing equations from data, providing a robust and reliable framework for future applications in robotic control and system analysis.