Research On Phase Optimization Of Relay Network Assisted By Intelligent Reflecting Surface

The traditional relay network has efficient signal processing capabilities,but it suffers from high circuit costs and power consumption.Intelligent reflecting surface(IRS)is a green wireless reflection technology with the advantages of low cost and low energy consumption,and can intelligently regulate the wireless environment to enhance signal transmission performance.The organic combination of the two can effectively complement each other’s strengths and weaknesses,achieve a good balance between circuit cost and rate enhancement,and provide new solution methods for the next generation of green wireless communication.This thesis focuses on the optimization and design of IRS phase and relay forwarding beamforming for IRS-assisted relay wireless networks.The main research content and innovations are as follows:1)In order to improve the rate of the IRS-assisted one-way amplify-and-forward relay network and maximize the signal-to-noise ratio,an alternating optimization algorithm for maximizing signal to noise ratio(Max-SNR-AO)is proposed.The proposed algorithm solves the phase shift(PS)matrix in the first time slot by the Lagrange function,calculates the PS matrix in the section time slot by Rayleigh quotient,and convert the amplification forwarding beamforming matrix into a convex problem by the CharnesCooper transformation.Simulation results show that the proposed Max-SNR-AO algorithm can significantly improve the rate compared with random phase.For example,the performance gain is up to 50% when the number of IRS reflecting elements is 512.2)In order to improve the system sum rate of the IRS-assisted two-way decode-andforward relay network,three IRS phase configuration algorithms are proposed:Maximizing received power sum plus eigenvalue decomposition(Max-RPS-EVD),maximizing minimum rate(Max-Min-R),and maximizing sum rate plus general power iteration(Max-SR-GPI).Among them,Max-RPS-EVD has the lowest computational complexity,and Max-SR-GPI has the best performance.Simulation results show that the sum rate of Max-SR-GPI is significantly better than Max-RPS-EVD,Max-Min-R,and random phase.In particular,when the number of IRS reflecting elements is 512,the proposed Max-SR-GPI can harvest a 53% sum rate gain compared with random phase.