Analysis Of Hierarchical Precoding Schemes For Massive MIMO Systems

Massive multiple-input and multiple-output (MIMO) has been regarded as one of the most promising technologies for the 5G system, it also be recognized as an essential extension of traditional MIMO technology.Intuitively, for the massive MIMO system, hundreds of antenna elements are equipped at the transceiver side to simultaneously serve dozens of users.However, tremendous antennas deployment provides with abundant degree of freedom but higher computational complexity and more serious user interference. To exploit the full potential of a massive MIMO system, one must leverage the precoding technique. Precoding is a mature signal pre-processing operation to greatly eliminate the user interference by the known channel state information (CSI). Unfortunately, traditional precoding techniques cannot be "transplant" directly to massive MIMO systems because of the intolerant computational complexity.In this context, this paper presents an exhaustive analysis on innovatively hierarchical precoding designs for massive MIMO systems. It includes a comparative analysis on the existing hierarchical precoding designs along with its performance analysis. In addition, this paper also proposes a hybrid precoding strategy with optimization objective of the weighted minimum mean square errors. The proposed algorithm embraces economic hardware cost and reduced computation complexity. The main contributions are as follows:First, the antenna configuration and antenna structure of massive MIMO systems are classified and well analyzed in the second chapter. Meanwhile,some channel models, which are more suitable for massive MIMO systems,are elaborated. Besides, as the hierarchical precoding technology is an extension of the traditional precoding,this paper also briefly describes four traditional linear precoding schemes. Then an exhaustive analysis on hierarchical precoding under the distinctive antenna structures are presented, meanwhile, both precoding performance gains and hardware complexity are studied. Our simulation results also reveal that precoding performance is still subject to hardware complexity. Furthermore, there is a tradeoff between precoding performance attained and hardware cost imposed. We also provide a guideline on antenna structure selection to achieve a better balance between the performance and cost.According to the former analysis, we can also concluded that full-array hybrid antenna structure obtains a decent precoding performance with a great hardware cost reduction. For this reason,a hybrid precoding strategy based on massive MU-MIMO system is proposed. The algorithm is mainly divided into the optimal precoding solution, "coarse" solution to the analog precoding and "precise" solution to obtain digital precoding. Both precoding algorithm complexity and convergence are verified. The proposed precoding minimizes weighted minimum mean square errors,and greatly reduces the hardware cost and obtains the near optimal performance gain.