Study Of Channel Non-reciprocity Compensation Correction Model And Antenna Selection For Optical Large MIMO System

Free space optical(FSO) communication has a lot of advantages than RF communication,such as high-speed optical bandwidth, without spectrum licensing, resistance to electromagnetic interference and so on. Large-scale MIMO technology as one of the key technology of 5G, can significantly improve system capacity and energy efficiency. When large-scale MIMO technology is applied to FSO system, it can not only effectively resist the harmful effects of atmospheric turbulence, but also overcome the pilot pollution in large-scale MIMO RF domain. The validity and reliability of the whole system are improved greatly,therefore, this paper studies optical large MIMO system.There exist two key problems in optical large MIMO system, one is the channel non-reciprocity between uplink and downlink, the other is the huge number and dense deployment of antennas. Although there also exist channel non-reciprocity in RF domain and it makes the CSI imperfect under TDD operation in large MIMO system. But in FSO, due to the presence of atmospheric turbulence, the channel non-reciprocity problem become more complex and can not be ignored. On the other hand, the huge number of antennas make the energy consumption and hardware cost of optical large system increase greatly, the dense deployment of antennas make the diversity or the reuse performance of the system reduce because there exists spatial correlation. Therefore, to study and put forward some effective antenna selection algorithms is also imperative. This paper exploits a single cell multi-user optical large MIMO system, aiming at the above two problems, study the channel nonreciprocity compensation correction model and antenna selection respectively. The main work of this paper is as follows:(1) We study the channel non-reciprocal compensation correction model in optical large MIMO system. First of all, we establish the system model, and establish the channel model of our system based on laser propagation in the atmosphere considering atmospheric path loss,beam geometry spread and beam intensity scintillation these three factors. Then, we study theinfluence factors of the channel non-reciprocity between uplink and downlink, including deriving the beam intensity attenuation and beam spread difference when the laser beam pass through atmospheric turbulence in the inclined path, and establishing a universal spatial correlation model to different intensity atmospheric turbulence in view of the spatial correlation difference between uplink and downlink. Finally, based on the system model and channel model we have established, and synthesizing the multiple influence factors of channel non-reciprocity, we propose the channel non-reciprocity compensation correction model for optical large MIMO system. Simulation results prove that the compensation correction model apply to different intensity atmospheric turbulence.(2) We study antenna selection in optical large MIMO system. First, aiming at the optical large MIMO system with antenna selection, we derive the approximate distribution of its channel mutual information, and prove that there exist channel hardening phenomenon in our system, the channel hardening phenomenon is more and more obvious as the total antenna number N becomes greater. Then, based on the definition of the energy efficiency of our system, aiming at the different conditions that the relative size of circuit power consumption and transmitted power are different, we study the effect of the transmitted power and antenna number on the energy efficiency of our system respectively. Finally, from the theory and practice two aspects, we study the antenna selection algorithm respectively with the goal of maximizing energy efficiency and maximizing interruption energy efficiency. We derive the selected number of antenna based on energy efficiency maximization, and derive the system interrupt average mutual information based on the Qo S standard of actual LTE system, and put forward the M_BSA antenna selection algorithm satisfying the interrupt energy efficiency maximization.