Physics-informed neural networks for one-dimensional sound field predictions with parameterized sources and impedance boundaries

TL;DR

This paper introduces a physics-informed neural network (PINN) approach for 1D sound field prediction that efficiently models dynamic sources and impedance boundaries, offering a promising step toward realistic 3D acoustic simulations.

Contribution

It presents a novel PINN model that learns a compact surrogate for 1D acoustics with parameterized sources and boundaries, addressing dynamic scene challenges.

Findings

Achieves relative mean errors below 2%/0.2 dB

Models parameterized moving Gaussian sources effectively

Lays groundwork for extending PINNs to 3D acoustics

Abstract

Realistic sound is essential in virtual environments, such as computer games and mixed reality. Efficient and accurate numerical methods for pre-calculating acoustics have been developed over the last decade; however, pre-calculating acoustics makes handling dynamic scenes with moving sources challenging, requiring intractable memory storage. A physics-informed neural network (PINN) method in 1D is presented, which learns a compact and efficient surrogate model with parameterized moving Gaussian sources and impedance boundaries, and satisfies a system of coupled equations. The model shows relative mean errors below 2%/0.2 dB and proposes a first step in developing PINNs for realistic 3D scenes.