Collaborative Research: NeTS: Medium: Design and Algorithms for Active Reconfigurable Intelligent Surfaces for sub-THz Wireless Networks

The tremendous growth of wireless data traffic over the past decades is expected to accelerate even more in future due to increasing demands for high-speed wireless connectivity, ubiquitous network access, and end-user experience. Sub-terahertz (THz) communications, defined as above 100 GHz, are envisioned as a key technology to enable the needed wireless terabit-per-second links by leveraging the hundreds of gigahertz bandwidths available at sub-THz bands. A major challenge in sub-THz bands, caused by higher propagation loss with increasing frequencies, is the limited communication distance. An emerging technology that promises to improve wireless coverage is the active reconfigurable intelligent surface (active-RIS) that consumes low power and provides efficient control of the reflected signals in both phases and amplification. Realizing this potential will require substantial research in hardware design and prototyping of wideband RIS operating above 100 GHz, as well as novel communication and network algorithms for active-RIS-aided wideband systems, together with experimental evaluation and validation of such unique sub-THz networks with active RIS.This project focuses on the 142 GHz frequency band as a front-runner for the first sixth-generation (6G) spectrum to be allocated above 100 GHz and a top choice for future Wi-Fi spectrum allocations in the years to come. The project consists of three intertwined thrusts. The first thrust is to design and prototype a wideband liquid crystal-based RIS with a wide angular range of tunable reflection operating at 142 GHz. Starting with a design for passive RIS as the proof-of-concept at this high frequency, an active RIS design will then be realized using amplifier-integrated LC-based substrate-integrated waveguide, enabling high tunability for each RIS element. The second thrust is to design robust and efficient algorithms for optimal control of the active RIS coefficients including frequency-dependent phase shift and amplitude amplification. Novel algorithms leveraging unsupervised graph neural networks and reinforcement learning will be used to capture the underlying network interaction and to provide strong scalability and generalizability. The third thrust is to perform extensive validation using the NSF-funded open-source ray-tracing simulation tool “NYURay” for active-RIS-aided sub-THz channel simulations. In addition, the prototyped passive and active RISs will be used to conduct on-site wireless propagation measurements utilizing the wideband sliding correlation channel sounder to create a site-specific hybrid channel model for RIS-aided communication. Through various education and outreach activities to broaden participation in computing, this project will foster knowledge sharing and contribute to industry and regulatory advancements in THz communications.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.