Neural-Driven Control of RIS in 6G Networks: A GoSimRIS and xApp-Based Framework

In this letter, we propose an O-RAN-based framework for Reconfigurable Intelligent Surfaces (RIS) control in 6G. The key objective is to enable the development of RIS control algorithms as xApps running at the Real-time Intelligent Controller (RIC) of Open RAN (O-RAN). To validate the proposed framework, we developed a Golang-based RIS simulator, GoSimRIS, intended to mimic and examine RIS behavior in various environmental scenarios. The simulator is linked with the RIC via a specialized Service Model (SM) devised in this work, namely E2SM RIS, which allows the design of xApps that dynamically optimize RIS coefficients by computing the ideal phase shifts and applying them in real-time to maximize network performance using channel information that is retrieved from the GoSimRIS environment. Finally, we introduce an ML-based RIS control mechanism that runs as an xApp using only the positions of the transmitter (Tx) and receiver (Rx) and the presence of Line-of-Sight (LOS) conditions, which corresponds to a realistic indoor scenario in 6G such industry 4.0.