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

This paper presented a novel device-free localization (DFL) framework based on distributed antenna networks (DANs), targeting integrated sensing and communication (ISAC) in future 6G radio access networks (RANs). 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 improve localization accuracy. 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 systems for accurate, robust, and scalable localization.