Research On Effective Capacity Analysis And Optimization Of RIS-aided Wireless Communication System

With the global deployment of 5G networks,in order to meet people’s ever-changing communication requirements and realize the real internet of everything,the research on 6G networks is imperative.It is expected to provide more stringent index requirements than 5G,such as ultra-high spectral efficiency(SE)and energy efficiency(EE),extremely high reliability and extremely low latency.In recent years,reconfigurable intelligent surfaces(RIS)technology is proposed as a new breakthrough to implement the next generation wireless networks,which will greatly boost the SE of wireless networks and the users’ quality of service(QoS).However,existing physical layer channel models do not explicitly take into account link-layer QoS metrics.In order to promote the efficient support of QoS in the future wireless networks,effective capacity(EC)is introduced in this thesis to measure the performance of RIS-aided wireless networks under specific delay QoS constraints.Therefore,this thesis focuses on analyzing the EC performance of RIS-aided networks in different scenarios,and further studies the phase-shift matrix design problem of RIS based on maximizing EC.Firstly,in order to quantify the impact of delay on system capacity,the EC performance of RIS-aided single-user downlink single-input-single-output(SISO)wireless networks is studied for Nakagami-m and Rician fading channels.The closed-form expressions for EC are derived under constant power transmission strategy,and the asymptotic expression in the high signal-to-noise ratio(SNR)region and analytical upper bounds are also obtained.Moreover,the closed-form expressions and the asymptotic results of EC are derived for different power and adaptive transmission strategies at the base station(BS),i.e.,water-filling algorithm and delay QoS driven power adaptive transmission strategy.On this basis,simulations results are provided to verify the accuracy of the derived expressions.The results show that increasing the number of RIS reflecting elements can effectively improve the EC of the system.In addition,the impacts of different channel conditions and adaptive transmission strategies on the EC performance of RIS-aided networks are also revealed.The obtained results provide a theoretical basis for the design and optimization of RIS-aided delay-sensitive networks.Secondly,in order to meet the ultra-reliable and low-latency communication(URLLC)requirements of mission-critical applications in Internet of Things(Io T)networks,the finite blocklength transmission scheme is considered.Based on this scenario,the EC performance of RIS-aided wireless networks in the finite blocklength regime is evaluated under the influences of delay QoS constraints and transmission error probability.The first-order statistical characteristic of the instantaneous SNR at the receiver of RIS-aided network is approximated by using moment-matching approach,then the approximate closed-form expression,the asymptotic expression in high SNR regime,analytical upper bound and low bound for EC are derived respectively.Numerical simulations are performed to verify the accuracy of the theoretical results,and the impacts of different system parameters on the EC are further analyzed.The results show that increasing the number of RIS reflecting elements or average SNR can improve the system EC within a certain range,whereas it will eventually approach the upper bound of system performance.Furthermore,there is a unique transmission error probability that maximizes EC.The theoretical results are useful for the design of RIS-aided URLLC networks.Finally,based on the analysis of the above EC performance,the RIS-aided single user multiple-input single-output(MISO)network with delay QoS constraints is further considered.Aiming at maximizing the effective capacity,the active beamforming vector at BS and the phase shift matrix at RIS are designed.In this thesis,it is assumed that maximum ratio transmission(MRT)scheme is adopted at BS.The phase shift matrix at RIS is designed based on the long-term channel state information(CSI)and stochastic gradient descent(SGD)algorithm is proposed to optimize the phase shift vector at RIS.Simulation results show that compared with the random phase shift configuration scheme,the proposed algorithm can effectively improve EC performance,and that a greater performance gain can be obtained for more RIS reflecting elements.In addition,the effects of the delay QoS exponent,the number of antennas at BS and other parameters on EC are revealed,which provides useful insights for the design of RIS-aided multi-antennas networks.