Fast Screening Coding Sequences for Programmable-Metasurface Reverberation Chambers Using a CNN-GA Framework
Year: 2026
Type: article
Abstract: Programmable metasurfaces provide a flexible electronic stirring approach for reverberation chambers (RCs). However, the selection of their coding states still largely relies on time-consuming full-wave simulations or measurement-based optimization. This letter presents a fast coding-sequence screening framework for programmable-metasurface RCs. The metasurface stirrer is first approximated as an array of electric dipole sources, and the chamber field is evaluated using a periodic Green's function. The resulting solver is then used to generate training data for a convolutional neural network (CNN), which serves as a fast field predictor. By embedding the trained CNN into a genetic algorithm (GA), coding sequences with low field correlation can be efficiently identified. A 1-bit programmable metasurface stirrer is used for proof-of-concept validation. The proposed method agrees well with full-wave simulations, reduces the field-evaluation time from tens of seconds to milliseconds, and yields coding-sequence sets with better field uniformity than random coding. These results indicate that the proposed CNN-GA framework is an efficient tool for rapid coding design in programmable-metasurface RCs.
Institution: Nanjing University of Aeronautics and Astronautics
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $2,645
OpenAlex ID: W7204969646
