Deep Learning Enables Accurate Sound Redistribution via Nonlocal Metasurfaces

Abstract

Conventional acoustic metasurfaces are constructed with gradiently "local" phase shift profiles provided by subunits. The local strategy implies the ignorance of the mutual coupling between subunits, which limits the efficiency of targeted sound manipulation, especially in complex environments. By taking into account the "nonlocal" interaction among subunits, nonlocal metasurface offers an opportunity for accurate control of sound propagation, but the requirement of the consideration of gathering coupling among all subunits, not just the nearest-neighbor coupling, greatly increases the complexity of the system and therefore hinders the explorations of functionalities of nonlocal metasurfaces. In this work, empowered by deep learning algorithms, the complex gathering coupling can be learned efficiently from the preset dataset so that the functionalities of nonlocal metasurfaces can be significantly uncovered. As an example, we demonstrate that nonlocal metasurfaces, which can redirect an incident wave into multi-channel reflections with arbitrary energy ratios, can be accurately predicted by deep learning algorithms. Compared to the theory, the relative error of the energy ratios is less than 1%. Furthermore, experiments witness three-channel reflection with three types of energy ratios of (1, 0, 0), (1/2, 0, 1/2), and (1/3, 1/3, 1/3), proving the validity of the deep learning enabled nonlocal metasurfaces. Our work might blaze a new trail in the design of acoustic functional devices, especially for the cases containing complex wave-matter interactions.

Sound‐Speed Modifying Acoustic Metasurfaces for Acoustic Holography

“…[74] Here, solid/ fluid-mechanical coupling via local evanescent fields and within the solid elements themselves can impact the uniformity of the acoustic field in the far field in structures that especially subject to these effects and can take advantage of this. Accordingly, design methods for nonlocal acoustic metasurfaces including coupled structures [74][75][76] and deep learning [77] have been reported. Our structures, however, incorporate relatively simple longitudinal elements, and while a local evanescent field evolves at pillar extrema (as shown in Figure S5 in the Supporting Information), this does not impact the acoustic field in the far-field, and where simulation results incorporate solid-mechanical coupling effects without strong evidence for the impact of evanescent fields and mechanical coupling through the metasurface structures.…”

Tunable, reconfigurable, and programmable acoustic metasurfaces: A review

“…Thanks to their thinness which is usually in the sub-wavelength, AMs have added value and functionalities in comparison with other acoustic metamaterials with small footprint (Xie et al, 2014;Cheng et al, 2015;Zhao et al, 2017;Assouar et al, 2018;Quan and Alu, 2019). Numerous exotic acoustic phenomena such as sound cloaking (Faure et al, 2016;Ma et al, 2019;Fan et al, 2020;Zhou et al, 2020), sound splitting (Zhai et al, 2018;Ding et al, 2021), sound absorption (Ma et al, 2014;Song et al, 2019;Liu et al, 2021;Li et al, 2022a;Guo et al, 2022), anomalous reflection or refraction (Diaz-Rubio and Tretyakov, 2017;Li et al, 2019a;Zhu and Lau, 2019;Li et al, 2020a;Chiang et al, 2020;Song et al, 2021), sound focusing (Zhu et al, 2016a;Lombard et al, 2022), one-way sound propagation (Zhu et al, 2015;Jiang et al, 2016), and medical ultrasound (Tian et al, 2017;Hu et al, 2022) have been proposed and demonstrated using AMs. AMs possess unusual features, including selective focusing and negative refraction, are enabled by the generalized Snell's law, which adds a new degree of freedom to control the behavior of transmitted or reflected waves by incorporating a lateral momentum (Yu et al, 2011) (see Figure 1).…”

Simple acoustic metagrating for perfect two- and three-beam splitting

“…Acoustic metasurfaces (Cummer et al, 2016;Assouar et al, 2018) are artificially designed structures composed of periodic subwavelength elements including groove structures , Helmholtz resonators (Li et al, 2018), labyrinthine structures (Xie et al, 2014), space coiling-up structures (Li et al, 2013;Chen et al, 2021), membranes (Ma et al, 2014), etc. In recent years, acoustic metasurfaces have attracted significant attention for their great potential applications in many fields attributed by their interesting and extraordinary acoustic properties (Zhao et al, 2022), such as anomalous reflection and refraction (Memoli et al, 2017;Zhu and Lau, 2019), acoustic focusing (Qi et al, 2017;Lombard et al, 2022), acoustic cloaking (Faure et al, 2016;Jin et al, 2019), sound absorption (Aurégan, 2018;, one-way acoustic propagation (Liang et al, 2010;Li et al, 2017;, beam splitting (Ding et al, 2021), etc. Particularly, acoustic beam splitters have attracted growing interest recently due to their applications in acoustic communication (Prada et al, 2007) and acoustic sensing (Dowling and Sabra, 2015) fields.…”

“…(Díaz-Rubio et al 2019) proposed a power flow-conformal acoustic beam splitter by manipulating the surface impedance distribution of metasurface to split the incident wave into two reflected beams propagating along two different directions. Beam splitting with arbitrary energy ratios can be realized by non-local grooved metasurfaces by using deep learning algorithms (Ding et al, 2021). Bianisotropic metasurface was developed for near-perfect arbitrary beam splitting by introducing self-induced surface waves into scattered field (Li et al, 2020).…”