High-Precision Intelligent Reflecting Surfaces-assisted Positioning Service in 5G Networks with Flexible Numerology

Accurate positioning is paramount for a wide array of location-based services (LBS) in fifth-generation (5G) wireless networks. Recent advances in 5G New Radio (NR) technology holds promise for very high-precision positioning services. Yet, challenges arise due to diverse types of numerology and massive connected devices. This paper presents a novel approach to improve positioning precision within a 5G NR framework with comb patterns on time-frequency resource mapping. We then formulate an optimization problem aimed at minimizing the maximum users' positioning error in an intelligent reflected surface (IRS)-assisted 5G network by controlling the user-anchor association, numerology-related selection, IRS's reflecting elements, privacy protection level, and transmit power. To address the non-convex nature of the underlying mixed-integer non-convex problem (MINLP), we propose an efficient algorithm that combines optimization, matching, and learning techniques. Through extensive numerical experiments, we demonstrate the effectiveness of our proposed algorithm in minimizing positioning errors compared to conventional methods.