A Study Of Energy Efficiency Optimization In Large-Scale MIMO Systems

Through transmitting and receiving data with large-scale antenna array, large-scale MIMO (multiple input multiple output) system can focus the energy of the beam on a small region of space, which results in the reduced transmitting and receiving power and the reduced interference to the other users, so it can achieve higher energy efficiency than traditional MIMO system, and it is an important candidate technology for 5G(5th Generation system for mobile communications). With the deepening of global warming, energy efficiency optimization becomes an important issue of scientific research and manufacture. In large-scale MIMO systems, due to the much more power consumed by the radio frequency chains which feed the huge antenna array, using all antennas to for data transmission will cause loss of energy efficiency, in the other hand, the complexity of large-scale MIMO systems give further optimization of energy efficiency more chances, so energy efficiency optimization of large-scale MIMO systems has been widely researched. This dissertation focuses on antenna selection method as well as the number of antennas, the number of users, and transmission power joint optimization problem, the main work of this dissertation are listed bellow.1. Low complexity antenna selection algorithms is studied. The number of antennas in large-scale MIMO (Multi-Input Multi-Output) systems is bigger than traditional MIMO systems, which makes antenna selection in large-scale MIMO systems more complex. Firstly, an antenna selection algorithm for uplink is introduced, it solves a MSE(Mean Squared Error) minimization problem, this problem can be converted into a sparse approximation problem, which can be solved through the existing sparse approximation methods. Based on sparse approximation method, an antenna selection algorithm for large-scale MIMO Multi-User systems is proposed in this dissertation. The proposed method takes the advantage of the diversity of the received signals form different antennas in large-scale antenna array, and do the antenna selection by solving a sparse approximation problem with the channel matrix as the measure dictionary. Since the method considers the correlation among the antennas antenna, it can achieve better performance than other low complexity antenna selection algorithms that ignores the correlation among antennas in large-scale MIMO systems. Simulation results show that the higher the correlation coefficient, the more obvious the performance gain of the proposed algorithm.2. Joint optimization of the number of antennas, number of users and transmit power on energy efficiency under ZF(Zero forcing) precoding is studied. Upon the analysis of ergodic energy efficiency with ZF precoding in large-scale MIMO systems, with regard to fractional optimization problem, the joint optimization algorithm of the number of antennas, transmit power, and the number of users and its degradations is given in this dissertation. In the optimization problem, transmission power and the number of antennas optimization can be converted into forms that can be solved efficiently, and the conditions of iteration optimization of this two are met. The number of users optimization can be solved by binary search algorithm. The joint optimization of the number of antennas, number of users and transmit power can also be simplified according to the properties of the expression of energy efficiency. The simulation results show that the proposed algorithm can realize an approximate maximum energy efficiency in both related and unrelated channels.3. Joint optimization of the number of antennas, number of users and transmit power on energy efficiency under MRT(Maximum Ratio Transmit) precoding is studied. Along with the idea in the joint optimization algorithm under ZF precoding, the ergodic energy efficiency under MRT precoding is analyzed, related proofs and deduction of the joint optimization algorithm under MRT precoding are given. Both multi-user systems and single-user systems are studied, for single-user systems, both the scenarios with and without CSI(Channel State Information) are considered.