Device-Free Localization Using Multi-Link MIMO Channels in Distributed Antenna Networks

Abstract

Targeting integrated sensing and communication (ISAC) in future 6G radio access networks (RANs), this paper presents a novel device-free localization (DFL) framework based on distributed antenna networks (DANs). In the proposed approach, radio tomographic imaging (RTI) leverages the spatial and temporal diversity of multi-link multiple-input multiple-output (MIMO) channels in DANs to achieve accurate localization. Furthermore, a prototype system was developed using software-defined radios (SDRs) operating in the sub-6 GHz band, and comprehensive evaluations were conducted under indoor conditions involving varying node densities and target types. The results demonstrate that the framework achieves sub-meter localization accuracy in most scenarios and maintains robust performance under complex multipath environments. In addition, the use of Bayesian optimization to fine-tune key parameters, such as sparsity and path thickness, led to significant improvements in image reconstruction quality and target estimation accuracy. These results demonstrate the feasibility and effectiveness of DAN-based DFL as a scalable and infrastructure-compatible ISAC solution, capable of delivering accurate, passive localization without dedicated sensing hardware.