Indoor Localization Using Dynamic DRSS Model in 5G System

Abstract

Received signal strength (RSS)-based localization methods are widely used in indoor positioning scenarios within 5G systems due to their cost-effectiveness and broad device compatibility. However, the path loss exponent (PLE) in the path loss model is highly sensitive to the localization environment, and precisely measuring the reference signal received power (RSRP) at the reference point remains challenging in practice. Consequently, in different localization application scenarios, continuous measurement and adjustment of the RSRP at the reference point and the PLE are required. Otherwise, the localization accuracy will be degraded. In this article, we first employ a dynamic difference of RSS (DRSS) model to eliminate the impact of RSRP measurement errors at the reference point. The model also addresses variations in PLE at different locations within the same localization scenario, as well as dynamic changes in PLE within the environment. Subsequently, a localization coordinate adjudicator is proposed to iteratively update the UE position and determine the optimal PLE for the current UE. Finally, under the optimal PLE, the UE’s localization coordinates are obtained using a genetic algorithm with a dynamic elite retention mechanism. Experimental validation was performed using both publicly available 5G simulation datasets and real-world data. The results show that the proposed dynamic DRSS model achieves a root mean square error (RMSE) of 2.44 m, outperforming existing techniques by 29%.