Super-resolution Acoustic Diagnostics Through Dynamic Wavelength Down-conversion

Acoustic diagnostics - using sound to understand objects, materials, and environments - plays an important role in many areas of broad societal impact, including biotechnology, medical imaging, ocean monitoring, and autonomous robotics. Yet current acoustic diagnostic systems still face major limitations. Many rely on electronically synchronized arrays of acoustic sensors and transmitters, which can restrict image resolution while increasing power consumption and system complexity. This project will create a new approach to acoustic diagnostics based on a process called wavelength downconversion, which converts sound-based information into a different wave that can produce much higher resolution images than conventional methods. The project will also advance the concept of “AI-ready diagnostics.” Rather than requiring new artificial intelligence (AI) algorithms to interpret complex, unfamiliar acoustic signals, the method in this project produces acoustic representations that resemble familiar photographic images. These images can then be analyzed directly using mature and powerful computer-vision tools developed and refined over more than a decade. In addition to its research goals, the project aims to strengthen acoustics education through a summer program for undergraduate students from partner institutions, modernization of an advanced acoustics course, and outreach workshops designed to increase public understanding of acoustics-based technologies. The goal of this project is to establish a new diagnostics framework in which acoustic fields scattered or generated by an environment, object, or medium are measured over an aperture, converted into a coherent field with a much shorter wavelength, and focused into a photographic-like image using a lens system. One possible implementation used for illustration purposes consists of a metasurface composed of independent unit cells that locally convert a long-wavelength acoustic field into a coherent, shorter-wavelength optical field. In this implementation, an off-the-shelf camera serves as the lens-based image-formation device. The resulting photo-like image can be analyzed directly using existing AI algorithms for visual image processing. Preliminary analysis indicates that this approach can achieve super-resolution - defined here as resolution beyond the conventional acoustic diffraction limit - while consuming very little power. This project will determine the fundamental limits on resolution and bandwidth for wavelength-downconversion-based diagnostics and quantify its performance in anisotropic and inhomogeneous media, including biologically relevant environments such as tissue. It is expected that carefully designed wavelength-downconversion systems can achieve resolutions orders of magnitude higher than those predicted by classical acoustic diffraction theory while maintaining broad operational bandwidth. This project will establish the potential of this diagnostic method for strategically important applications, including biotechnology research and biomedical imaging. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.