Research On User Localization Algorithms For Reconfigurable Intelligent Surface Aided MIMO Systems

With the accelerated construction of 5G mobile communication networks,the "Internet of Everything" is gradually becoming a reality,and the rapid growth of data traffic has placed higher demands on mobile communication systems.In recent years,Reconfigurable Intelligent Surface(RIS)has been attracting widespread attention from academia and industry because of its ability to flexibly manipulate the electromagnetic properties of the channel environment.RIS intelligently reconfigures the wireless propagation environment through directional reflection of incident electromagnetic waves,significantly reduces the energy consumption of electromagnetic wave propagation and significantly improves spectrum efficiency,enabling not only the ubiquitous communication,but also the sensing and localization services for users and devices.However,the integration of user localization and RIS in wireless communication systems still faces many urgent problems,such as high complexity of channel modeling,high precision localization of mobile users with strong time dependence,and difficulty in efficient design of RIS phases.This thesis addresses the above problems and investigates the high-precision localization algorithm in RIS-assisted Multiple-Input Multiple-Output(MIMO)systems,and the main research contents are summarized as follows:To address the location problem of non-time-varying users,this thesis focuses on a MIMO system composed of multiple RIS-assisted multi-antenna base stations and multiantenna users deployed in a distributed manner.First,based on the scenario that the direct transmission path between the base station and the user is blocked,RIS is used as a passive relay to build a cascade channel consisting of base station,RIS and non-timevarying users.From the non-time-varying characteristics of the user’s location,an algorithm is proposed to obtain user location information based on the Angle of Arrival(AOA)of the RIS reflected signal at the user’s antenna and the non-time-varying geometric relationship constructed by the base station and RIS with known specific deployment location.Secondly,the base station beam assignment design and RIS phase design are carried out based on the Cramér-Rao Lower Bound(CRLB)of the non-timevarying user’s location estimation to achieve the purpose of optimizing the localization algorithm and improving the performance of the localization system,respectively.The simulation results show that the accuracy of AOA estimation and position coordinate estimation included in the algorithm is significantly improved when the number of RIS deployment,the number of sampling snapshots,the number of samples and the signal-tonoise ratio of the localization algorithm are higher.To address the localization problem of users with time-varying locations,this thesis considers the temporal correlation of the location of mobile users based on the study of non-time-varying user localization.A Markov probability transfer model for time-varying user location movement is established,and an online user location estimation algorithm is proposed based on the information transfer principle,and the transfer of information between two modules,AOA and location coordinate estimation,is analyzed by factor graphs.The algorithm approximates the posterior distribution of AOA estimation with a von Mises distribution and uses it to calculate the estimation of user location coordinates along a Markov chain,and derives the basic performance of the CRLB for this scenario to measure the location problem for time-varying users.To improve the system performance,the beamforming of the base station and the RIS phases are jointly designed by minimizing the CRLB.Simulation results show that an online localization algorithm proposed in this thesis yields the performance of the AOA and the user’s coordinate estimation,both of which are close to the CRLB at high signal-to-noise ratios.