MimoLoc: Indoor Localization With Assistance of Microwave Reflection of Downlink Signal in Sub-6G MIMO Networks

Compared to the localization at the network side, indoor localization with high accuracy using a downlink signal has many advantages, such as preserving location privacy and easing the network’s computation burden and overhead data. In this article, we propose MimoLoc, an accurate indoor localization system that leverages microwave reflection in sub-6G multiple-input multiple-output (MIMO) networks. MimoLoc utilizes both the direct and reflection paths of the downlink signal to locate the terminal device, effectively mitigating phase errors caused by signal synchronization error between the access point (AP) and the terminal device. To accomplish localization, we establish a triangle-based localization model that integrates both a reflection path and a direct path. Subsequently, we derive the Cramér-Rao lower bound (CRLB) of the proposed localization model and introduce the shape factor (SF) to reveal the correlation between localization accuracy and the shape of the multipath triangle. In contrast to a uniform linear array (ULA), our investigation reveals that using a uniform circular array (UCA) enhances the attainability of reflection paths associated with reduced localization errors. As a result, we use UCAs in conjunction with sparse signal recovery techniques to extract weak reflection paths and the direct path. To validate the efficacy of MimoLoc, we develop a prototype using software-defined radio devices and design the signal structure for the MIMO channel estimation. Extensive experiments conducted in various indoor environments demonstrate that MimoLoc achieves an accuracy of 0.58 m using only a single AP.