Indoor Signal Focusing with Deep Learning Designed Reconfigurable Intelligent Surfaces

Abstract

Reconfigurable Intelligent Surfaces (RISs) comprised of tunable unit elements have been recently considered in indoor communication environments for focusing signal reflections to intended user locations. However, the current proofs of concept require complex operations for the RIS configuration, which are mainly realized via wired control connections. In this paper, we present a deep learning method for efficient online wireless configuration of RISs when deployed in indoor communication environments. According to the proposed method, a database of coordinate fingerprints is implemented during an offline training phase. This fingerprinting database is used to train the weights and bias of a properly designed Deep Neural Network (DNN), whose role is to unveil the mapping between the measured coordinate information at a user location and the configuration of the RIS's unit cells that maximizes this user's received signal strength. During the online phase of the presented method, the trained DNN is fed with the measured position information at the target user to output the optimal phase configurations of the RIS for signal power focusing on this intended location. Our realistic simulation results using ray tracing on a three dimensional indoor environment demonstrate that the proposed DNN-based configuration method exhibits its merits for all considered cases, and effectively increases the achievable throughput at the target user location.Index Terms-Reconfigurable intelligent surface, deep neural networks, channel state information, fingerprinting, indoor signal focusing, location information.

Unsupervised Learning for Passive Beamforming

“…In this section, simulation results are provided to demonstrate the performance of the proposed unsupervised learning approach. 2 We consider a similar indoor scenario as [14], where all the channels are modeled by independent Rayleigh small-scale fading, and the path loss in dB is computed as 20.4 log 10 (d/d ref ) [19], with d being the distance between transmitter and receiver in meters and d ref = 1m denoting the reference distance. As illustrated in Fig .1, the distance from AP to RIS is denoted by d AR , while the distance from AP to user and from RIS to user can be computed as d AU = d 2 0 + d 2 1 and d RU = (d AR − d 0 ) 2 + d 2 1 respectively, where d 1 is the vertical distance from user to the horizontal connection line of AP and RIS while d 0 is the distance from AP to the intersection.…”

Smart Radio Environments Empowered by Reconfigurable Intelligent Surfaces: How It Works, State of Research, and The Road Ahead

“…In [267], the authors employ an artificial neural network to maximize the received power in a RIS-based network. The downlink of a multi-user MISO system is considered, and the deep learning framework developed in [27] is applied.…”

A Prospective Look: Key Enabling Technologies, Applications and Open Research Topics in 6G Networks

“…This would support self-organization and automation of all metasurface functions, including maintenance, management, and operational tasks. A recent research study proposed a deep learning (DL)-based ML approach to achieve signal focusing through learning the mapping between the estimated channel state information (CSI) at a user location and the optimal configuration of the metasurface's unit cell [145]. Furthermore, adaptive control and coordination of multiple metasurfaces in programmable wireless environments was demonstrated for a set of users through the application of neural networks [146].…”

“…Researchers also explored the use of metasurfaces in wireless power transfer (WPT) [158]- [160]. The role of machine learning in controlling the functionalities of metasurfaces to actively improve the coverage of the highly dynamic indoor environments is analyzed in [161]- [163]. The aforementioned state-of-the-art is summarized in Table 6.…”

“…Modulation [149]- [153] Beamforming [154]- [157] Wireless power transfer (WPT) [158]- [160] Machine learning [161]- [163]…”