Research On Reconfigurable Intelligent Surface-Assisted Near-Field Localization

With the emergence of novel applications such as autonomous driving and augmented reality,the requirements for positioning accuracy and reliability have become increasingly important in the fifth and sixth generation mobile communication systems.The adoption of new frequency bands at millimeter-wave and terahertz guarantees higher available bandwidths,corresponding to potentially higher positioning accuracies,but,at the same time,it opens up new challenges in terms of coverage and reliability because signals can be blocked by obstacles.A promising technology to tackle this problem is represented by the reconfigurable intelligent surface(RIS),which allows the realization of intelligent and controllable wireless channel environments by adjusting the amplitudes,phase shifts,and other information of reflected(refracted)signal signals.RIS-assisted localization has become one of the research hotspots in recent years.However,most current research is based on the far-field assumption,whereas the presence of large antenna arrays operating at high frequency bands highlights the need to take the near-field effect into consideration.Therefore,this thesis mainly focuses on RIS-assisted near-field localization.Firstly,this thesis presents a comprehensive introduction to the fundamental principle of RIS,followed by an analysis of the characteristics and the model of near-field signals.Subsequently,two classic wireless positioning technologies,namely direct positioning and two-step positioning,are summarized and reviewed.Furthermore,common performance evaluation metrics for wireless localization systems are introduced.Secondly,for a RIS-assisted multi-user system,this thesis presents a near-field direct localization scheme with users’ cooperation.The proposed scheme consists of two steps.Firstly,a direct localization algorithm based on reflected signals is introduced to obtain initial position information.Secondly,a joint maximum likelihood algorithm for reflection and cooperative signals is proposed to achieve enhanced estimation accuracy.In the second step,the Cramér-Rao lower bound(CRLB)is derived,and a joint optimization problem that involves optimizing power allocation among the access point(AP)and users,as well as beamforming at the RIS,is proposed to minimize the CRLB.To solve this problem,an alternating optimization(AO)method is utilized.Finally,simulation results demonstrate that the proposed scheme outperforms non-cooperative scheme in terms of localization performance and that it can adaptively allocate power to improve energy efficiency.Finally,for a localization system assisted by simultaneous transmitting and reflecting RIS(STAR-RIS),where the users are distributed on both sides of the STAR-RIS,a novel near-field two-step positioning method based on polar domain sparsity is proposed,which exhibits lower search complexity compared to the direct positioning method.To be specific,by applying the second-order Taylor expansion approximation to the near-field signals,the estimation problem of channel parameters,i.e.,angles and distances,is transformed into a sparse recovery problem,which can be solved by the orthogonal matching pursuit(OMP)algorithm.The position estimation is then obtained based on geometrical relationships between the channel parameters and user positions.Next,a joint optimization of power allocation and beamforming for reflection and refraction signals is proposed to obtain more accurate position estimations.Simulation results demonstrate that compared to the conventional far-field estimation scheme,the proposed near-field scheme has better channel parameter estimation performance.Moreover,the results validate the superiority of STAR-RIS over the ordinary RIS in the omnidirectional positioning scenario.