Channel Estimation And Over-the-air Computation Of MIMO Communication With Electromagnetic Reflection

The commercial use of the fifth-generation mobile communication system has become widespread,and the intelligent society is gradually becoming a reality.However,the key physical layer technology,millimeterwave multiple-input multiple-output(MIMO)technology,still has obstacles.On the one hand,millimeterwave communication is easily blocked and the path loss is serious,resulting in limited coverage.On the other hand,the traditional massive MIMO technology needs to be equipped with a massive radio frequency,which consumes a lot of power and costs.However,emerging applications such as digital twins,virtual reality have put forward higher requirements for high-quality services in the new generation of mobile communication systems.Reconfigurable Intelligent Surface(RIS)is a planar array composed of a large number of almost passive low-cost reflective elements.With the help of an intelligent controller,each element independently introduces a phase shift to the incident signal,thereby overcoming unfavorable propagation conditions.The transceiver design of the RIS-assisted system is largely determined by the acquisition of Channel State Information(CSI).Thus,it is particularly important to study the channel estimation of RIS-assisted MIMO communication systems.This thesis considers the channel estimation of single-user RIS-MIMO system and multi-user RIS-MIMO system,and studies the data aggregation calculation problem of the over-the-air computing system where RIS is involved.Firstly,this thesis focuses on single-user RIS-MIMO communication system,the second chapter studies the channel estimation problem of passive RIS-assisted single-user system under hybrid finite-precision Analog-to-Digital Converter(ADC)architecture.For the uplink of a single-user RIS-MIMO system,a highperformance channel estimation scheme is proposed,the closed-form solution of the optimal estimator in the sense of mean square error is derived,and the nonlinear quantization noise of the finite-precision ADC is effectively suppressed.The thesis first applies Bussang’s theorem to quantify the finite-precision ADC modeling,which makes the mathematical problem solvable.The thesis then uses the sparse characteristics of the millimeter wave channel in the angle domain to reshape the problem,reducing the complexity of the problem solving,and obtains the closed-form solution of the channel estimation.Finally,the thesis further studies the implementation process of the specific channel estimation scheme,including subarray division,analog estimator design,digital estimator design,cascaded channel estimation and original channel recovery.The simulation results verify the proposed channel estimation performs better than traditional solutions.Secondly,this thesis focuses on the multi-user RIS-MIMO communication system,the third chapter studies the channel estimation problem of passive RIS-assisted multi-user system.For multi-user communication systems,closed-form solutions for channel estimation of direct path channels and cascaded channels of each user are derived.For the direct path channel,each unit of the RIS is turned off,and the optimal direct path channel estimation in the sense of mean square error is derived through the uplink.The thesis then studies the selection strategy of typical users,and completes the measurement of cascaded channels by selecting typical users with better channel quality.Then,using the same characteristics of the channels between BS and RIS for each user,the switch of the RIS unit is further studied.And by solving the relative channel parameters of atypical users relative to typical users,the cooperative cascaded channel estimation in the sense of mean square error is obtained.The dissertation uses a phased channel estimation scheme to improve the performance of channel estimation by selecting typical users and assisting channel estimation for atypical users,and effectively reduce the problem of excessive pilot overhead for channel estimation.Finally,this thesis focuses on the over-the-air computing system based on RIS communication,the fourth chapter of the thesis studies the problem of minimizing the computational distortion in the over-the-air computing system of RIS communication.For CSI additional bounded uncertainty conditions and total power constraints,the design of the transceiver and RIS phase is jointly optimized from the perspective of worstcase robust design.For the complex non-convex problem in the case of CSI error bounded uncertainty in the system,the thesis first uses the Karush-Kuhn-Tucker(KKT)condition to transform the non-convex problem,and then solves the KKT multipliers to obtain a closed-form solution to the robust problem.The thesis then takes the closed-form solution into the optimization problem to optimize the transceiver parameters and RIS phase.A closed-form solution for the joint transceiver and RIS design is obtained by using alternating optimization.That is to say,the transceiver parameters are fixed first,and the optimal RIS phase is obtained,and then the optimal RIS phase is brought into the optimization problem to obtain the optimal transceiver parameters.The simulation results verify the effectiveness of the proposed optimal design scheme in reducing computational distortion.