Programmable wave-based analog computing machine: a metastructure that designs metastructures

Area of Science:

• Physics
• Materials Science
• Computer Engineering

Background:

• Metastructures offer a novel paradigm for wave-based analog computing.
• Existing approaches face limitations in speed, loss, and compactness.
• Electromagnetic wave-based computation presents opportunities for near-speed-of-light performance.

Purpose of the Study:

• To theoretically introduce and experimentally verify a reconfigurable metastructure for analog complex mathematical computations.
• To demonstrate the device's capability in solving both stationary and non-stationary iterative algorithms.
• To explore the potential of wave-based analog computing in compact and ultrafast devices.

Main Methods:

• Development of a reconfigurable metastructure utilizing radio frequency (RF)-based components.
• Theoretical introduction and experimental verification of the metastructure's computational capabilities.
• Application of the metastructure to matrix inversion, root finding (Newton's method), and inverse design (Lagrange multiplier method).

Main Results:

• Successful demonstration of analog matrix inversion using the reconfigurable metastructure.
• Experimental verification of solving non-stationary problems, including root finding and inverse design.
• The metastructure performed computations in a compact, ultrafast, and parallelized manner.

Conclusions:

• The reconfigurable metastructure successfully performs complex analog mathematical computations using electromagnetic waves.
• This platform opens avenues for general linear algebraic problems and beyond via wave-based analog computing.
• The developed device offers a compact, ultrafast, and parallelized approach to analog computation.