RIS-Aided Wireless Multiport Sensing: Multiplexed De-Embedding of an RIS-Programmable Over-the-Air Fixture

Abstract

Wireless multiport sensing aims to remotely retrieve the scattering matrix of a multiport device under test (DUT) connected to not-directly-accessible (NDA) antennas, based on scattering parameters measured with remotely located accessible antennas that couple over the air (OTA) to the NDA antennas. A central bottleneck is that the required number of accessible antennas grows with the number of unknowns in the DUT's scattering matrix. Here, we address this bottleneck by using a reconfigurable intelligent surface (RIS) to generate measurement diversity during the DUT characterization. We interpret the setup as measuring the DUT via an RIS-programmable OTA fixture. We first characterize the RIS-programmable OTA fixture using a specific known tunable load network to terminate the NDA antennas. Then, we reconstruct the DUT's scattering matrix by multiplexed de-embedding of the RIS-programmable OTA fixture, using measurements of the DUT acquired across multiple RIS configurations. Based on our system model formulated in terms of multiport-network theory, we quantify and maximize the diversity of our measurement sequence by optimizing the deployed ensemble of RIS configurations. We validate our approach experimentally at 2.45 GHz in a rich-scattering radio environment. We systematically examine the influence of the number of accessible antennas, the number of RIS configurations, and the optimization of the RIS configurations. For a SISO link (two accessible antennas), optimized RIS configurations reduce the median reconstruction mean-squared error for 100 DUT measurements by 45% from $1.57{\times}10^{-3}$ to $8.59{\times}10^{-4}$. These results demonstrate that RIS diversity can compensate for a limited number of accessible antennas, paving the way toward low-cost wireless multiport sensing for industrial, biomedical, and smart-environment RFID systems.