Solving a kind of inverse scattering problem of acoustic waves based on linear sampling method and neural network

Abstract

This paper explores the inverse scattering problem that reconstructs the obstacle shape with far-field information in the acoustic environment. Taking the planar wave as the incident wave, the authors put forward a shape reconstruction method for an impenetrable obstacle with a sound-soft boundary. Firstly, the linear sampling method (LSM) was performed to acquire the prior information about the obstacle shape. Next, a shape parameter inversion model was constructed based on neural network and gating thought, and it was denoted as the SPIMNNG. The model parameters were updated through self-learning. After that, the shape parameters of obstacle boundary were inversed according to the far-field information and the prior information about obstacle shape. On this basis, the boundary shape of the obstacle was reconstructed. The computing complexity of the SPIMNNG model was also calculated. Finally, several numerical experiments were carried out to verify the proposed method. The results show that our method can effectively inverse the shape parameters of obstacles.

Keywords

• Acoustic inversion
• Linear sampling method
• Neural network
• Gating thought
• Obstacle shape reconstruction