A joint Computer Vision and Reconfigurable Intelligent Meta-surface Approach for Interference Reduction in Beyond 5G Networks

Reconfigurable Intelligent Meta-surfaces (RIMs) are particular devices able to control and manipulate radio frequency wireless signals. This promising technology allows to improve the reliability of wireless networks, thanks to the capacity of reflecting the desired signals through appropriate phase shifts. The joint use of RIMs and Computer Vision (CV) technology is the main objective of this paper. This synergistic approach is used to correctly identify the specific configuration of a radiation pattern, to be used as input for computing optimal coding sequences of the RIM. Indeed, by the means of a CV algorithm it is possible to infer a connectivity graph related to a real scenario, where people is moving. The information about network nodes such as their distance, the relative position, etc. is used for feeding an intelligent logic, able to compute the optimal configuration for re-directing the signals towards a given receiver target node.Numerical results show the huge potentiality of this combined approach in terms of interference reduction. It has been observed that for high traffic load, it is possible to reduce the average interference in the network of 40%. Furthermore, an analysis including the positioning estimation error of the CV algorithm has been addressed, in order to consider how it affects the interference reduction. Results show that, even though there is an increasing effect of interference, when the error is accounted, the interference reduction impact is still important.