On identifying the non-linear dynamics of a hovercraft using an end-to-end deep learning approach

TL;DR

This paper introduces a novel deep learning approach to identify and control the non-linear dynamics of a custom hovercraft, demonstrating accurate predictions and effective closed-loop control on a real system.

Contribution

It presents an end-to-end deep learning method for modeling and controlling a hovercraft's non-linear dynamics directly from data, a novel application in this context.

Findings

Accurate non-linear dynamic models learned from data.

Effective position control demonstrated on the real hovercraft.

Successful closed-loop control performance achieved.

Abstract

We present the identification of the non-linear dynamics of a novel hovercraft design, employing end-to-end deep learning techniques. Our experimental setup consists of a hovercraft propelled by racing drone propellers mounted on a lightweight foam base, allowing it to float and be controlled freely on an air hockey table. We learn parametrized physics-inspired non-linear models directly from data trajectories, leveraging gradient-based optimization techniques prevalent in machine learning research. The chosen model structure allows us to control the position of the hovercraft precisely on the air hockey table. We then analyze the prediction performance and demonstrate the closed-loop control performance on the real system.