The Research Of Localization Technology Of Indoor Devices By Fusing Multipath Features

In recent years,with the rapid popularization of smart mobile devices and the fast growth of mobile Internet services,indoor localization based on wireless signals has been widely used in many fields such as security monitoring,indoor navigation,smart medical care,etc.However,the existing indoor wireless localization technologies do not fully exploit the multipath features of wireless signals and suffer from the following limitations: insufficient scalability under the scenario of limited antenna resources,poor reliability under the scenario of Non-Line-of-Sight(NLo S)propagation,and high hardware overhead under the scenario of large-scale deployment.To address these challenges,this thesis deeply explores and fuses the multipath features of wireless signals.It focuses on three aspects including single-antenna device localization,multi-antenna device localization,and passive device localization,and proposes a peer-to-peer localization scheme under the scenario of limited antenna resources,a reliable localization scheme under the scenario of NLo S propagation and a low-cost localization scheme under the scenario of large-scale deployment,respectively.The main contributions are summarized as follows:1、To address the challenge of insufficient scalability under the scenario of limited antenna resources,this thesis conducts research on peer-to-peer localization technology for single-antenna devices.Based on the core idea of actively controlling multipath signals,first,a high-precision distance estimation algorithm is designed to obtain accurate distance estimates between single-antenna devices by fusing Time of Fight and Doppler Frequency Shift of the target path using Kalman Filtering.Second,a location-retracing-based angle estimation algorithm is proposed to obtain accurate angle estimates between single-antenna devices by comparing the distance estimates of the direct path and controllable reflection path.Finally,the distance and angle estimates are fused to achieve peer-to-peer localization of single-antenna devices.Experimental results show that the proposed scheme achieves indoor single-antenna device localization with a median error of 0.88 m.2、To address the challenge of poor reliability under the scenario of NLo S propagation,this thesis conducts research on NLo S localization technology for multi-antenna devices.Based on the core idea of fusing the high-dimensional multipath features,first,a lowcomplexity high-dimensional parameter estimation algorithm is designed to separate the parameters corresponding to the reliable multipath simultaneously,such as Angle of Arrival,Angle of Departure,and Time of Flight,etc.Second,the geometric relationship between multipath high-dimensional features and target location and reflector location is modeled,and only the feature related to the target location is decoupled.Finally,a low-complexity iterative optimization algorithm is designed to eliminate the errors due to hardware imperfection and rotational uncertainty to achieve reliable NLo S localization of multi-antenna devices.The experimental results show that compared with the existing schemes,the proposed scheme improves 44% and 200% in the median accuracy and 90% accuracy,respectively.3、To address the challenge of large hardware overhead under the scenario of large-scale deployment,this thesis conducts research on low-cost localization technology for backscatter tags.Based on the core idea that the backscatter tags communicate by leveraging environmental radio signals,first,a highly sensitive multipath parameter estimation and backscattered signal recovery algorithm are designed to accurately estimate the backscattered signal with very low signal-to-interference-plus-noise ratio and the corresponding parameters.Second,the location estimation ambiguity of the tags under the underdetermined condition is analyzed and eliminated by fusing the motion information of the transmitter.Finally,based on particle swarm,low-cost tag localization is achieved.The experimental results show that the proposed scheme can achieve localization of multiple tags with a median error of 0.74 m.This thesis investigates localization technologies by fusing multipath features and verifies the performance of the proposed schemes by building a prototype experimental system,which provides a theoretical basis and technical support for the development and application of indoor wireless localization.