Cooperative Transmission Theory And Methods For Intelligent Reflecting Surface Aided Wireless Communications

The fundamental challenge for achieving ultra-reliable wireless communications arises from the time-varying wireless channels.Traditional approaches for tackling this challenge have been limited to transmission design at transceivers,which are unable to effectively deal with complex time-varying propagation environments.Intelligent reflecting surface(IRS)can achieve smart and reconfigurable wireless channels/radio propagation environment,and thus has become a key candidate technology for the 6th generation(6G)wireless communication systems.By properly adjusting its phase shifts,IRS can enhance the desired signal power at the receiver.However,due to the special architecture of IRS,IRS-assisted wireless communications faces many new challenges,mainly including: 1)how to acquire channel state information(CSI)with low training overhead;2)how to jointly design passive beamforming at the IRS and active beamforming at the base station(BS)with low complexity.Hence,focusing on the aforementioned challenges,this thesis conducts a very intensive study and obtains a number of innovative results as follows:First,a novel semi-passive elements-aided channel estimation framework is designed for IRSaided communications,where a small portion of IRS reflecting elements with the capability of processing the received signals are utilized to reduce training overhead.Specifically,the BS-IRS channel is estimated by applying the estimation of signal parameters via rotational invariance technique(ESPRIT),while the user-IRS channels are estimated by combining the use of total least square(TLS)ESPRIT and multiple signal classification(MUSIC)methods.The required training time of the proposed channel estimation scheme is irrelevant to the number of IRS reflecting elements,thus substantially reducing the training overhead.Moreover,our proposed scheme has great advantages over both the conventional compressed sensing(CS)-based channel estimation and cascaded channel estimation schemes in terms of channel estimation accuracy.Secondly,for the IRS-assisted single-user system,beamforming schemes under different CSI conditions are designed,including statistical CSI,partial CSI(i.e.,angle information),and imperfect CSI.First,to avoid high training overhead caused by the acquisition of instantaneous CSI,a statistical CSI based design framework is proposed,where statistical CSI is exploited for the joint design of BS transmit beamforming and IRS beamforming.The proposed beamforming schemes not only have much lower complexity,but also achieve similar performance as the benchmark algorithm requiring instantaneous CSI.Furthermore,an angle-domain design framework is proposed,where angle information is estimated with extremely low training overhead and then used for low-complexity beamforming design.Also,a closed-form expression for the achievable rate is derived.The analysis of the angle estimation error reveals the impacts of BS antenna number and IRS locations on the estimation accuracy,while the analysis of the achievable rate reveals the power gain brought by the use of partial CSI(i.e.,angle information)for beamforming design.In addition,considering the channel estimation error caused by location uncertainty,a robust beam design framework based on imperfect CSI is proposed.By jointly designing the active beamforming at the BS and passive beamforming at the IRS,we aim to minimize the transmit power subject to the worse-case quality of service(Qo S)constraint.By utilizing techniques of alternating optimization,Taylor expansion,S-Procedure and semi-definite relaxing(SDR),the joint optimization problem is transformed into a sequence of easy-to-solve semi-definite programming(SDP)sub-problems.The simulation results show that the proposed robust beamforming algorithm can effectively improve the reliability of user communication.Finally,the IRS-assisted single-user system is extended to IRS-assisted multi-user systems,including broadcast systems and multicast systems.For the multi-IRS-assisted broadcast system,a novel location-aided channel estimation and beamforming design framework is proposed,which effectively reduces the training overhead of channel estimation and the complexity of beamforming design.Furthermore,a closed-form expression of the achievable rate is derived,which reveals for the first time the impacts of location uncertainty,IRS deployment locations,and user locations on the achievable rate.Finally,an optimal power allocation scheme has been proposed to further improve the system performance.For the IRS-assisted multicast system,we consider a practical phase shift model with limited resolution,and propose a novel beam training-based design framework,where the beam training method achieves comparable performance as the exhaustive beam searching method but with much lower complexity and training overhead.Furthermore,the closed-form expression for the achievable rate is derived.In addition,for certain asymptotic scenario,closed-form solutions are obtained for the optimal power allocation scheme.