Research On Beamforming Technology For Distributed Massive MIMO Systems

As one of the key physical layer technologies in 5G era,massive MIMO has received extensive attention in various new scenarios in B5G and 6G era.Distributed massive MIMO is an important part of the future network technologies because of conveniences of flexible deployment,scalability and multi-point cooperation.Meanwhile,beamforming based on channel state information(CSI)can increase the received signal strength of users and suppress interference,which is a key research topic in massive MIMO systems.However,the computational complexity of the optimal beamforming design increases rapidly with the number of antennas in massive MIMO systems.In addition,the high overhead of global channel estimation and CSI exchange between transmission points in distributed deployment bring severe challenges to the distributed system.Therefore,we designed beamforming algorithms with low dependence on global instantaneous CSI in the following two distributed massive MIMO scenarios,by using the channel characteristics in different scenarios,exchanging statistical CSI and making full use of local instantaneous CSI.We give simulation results in different aspects.The main research contents include:Firstly,a distributed cross-link interference(CLI)suppression scheme based on beamforming is designed in the dynamic time division duplex(TDD)scenario.In the dynamic TDD scenario,the interfering base station(BS)using massive MIMO technology in the downlink generates strong CLI to the disturbed BS receiving in the uplink.Firstly,combined with the characteristics of the fixed position of the BSs and the relatively slow change speed of the channel between the BSs,the statistical CSI of CLI channel and eigenvalue decomposition are used to align the interference to the null space of the desired signal for pre-beamforming.Then,based on the zero-forcing criterion,the precoding matrix of the interfering station is designed by using the instantaneous CSI from the interfering BS to the users to suppress the inter-user interference.Finally,the downlink beamforming matrix of the interfering BS is formed by the pre-beamforming matrix and the precoding matrix.The proposed algorithm only needs to exchange statistical CSI between BSs,and does not require frequent channel estimation between BSs,and realizes distributed processing at the interfering station.The simulation results show that the proposed algorithm can effectively suppress the CLI between BSs at the cost of slightly reducing the downlink transmission capacity of the interfering station,and then greatly improve the uplink transmission capacity of the disturbed station.Secondly,we designed a beamforming scheme based on the association of intelligent reflecting surface(IRS)and user equipment(UE)in the virtual distributed massive MIMO system scenario composed of distributed IRSs and multiple BSs.In the scenario of distributed deployment of IRSs,the distances between each IRS and each user and the distances between each IRS and each BS are different,which are affected by the product-distance path loss model.In this paper,we derive the average signal to interference plus noise ratio(SINR)and design an IRSUE association scheme based on large scale fading parameters.After determining the IRS-UE association,we propose a distributed passive beamforming algorithm based on 1-bit feedback.By using the received signal strength to adjust the phase of the reflection elements,which does not require accurate CSI related to the IRSs to calculate the 1-bit feedback,the performance of the joint active and passive beamforming phase alignment scheme is gradually approached.The simulation results show that there is a significant difference in the average SINR between the user assisted by an IRS and the user without an IRS,and the converged 1-bit feedback algorithm can approach the performance of joint active and passive beamforming phase alignment,while reducing the requirement for high channel estimation accuracy.