AFC-CE Hybrid Precoding for Energy-Efficient mmWave MIMO Systems

Abstract

In this letter, an energy-efficient hybrid precoding method based on cross-entropy (CE) is proposed for the adaptive fully-connected (AFC) structure in millimeter wave massive multiple-input multiple-output systems, named the AFC-CE algorithm. By the nonlinear Tomlinson-Harashima precoding method, the non-convex spectrum efficiency maximization problem is transformed into a matrix residual minimization problem between the hybrid precoding matrix and the optimal precoding matrix. Then, using CE theory, the matrix residuals minimization problem is transformed into a probability distribution optimization problem of switching precoding matrix elements. Finally, the probability distribution parameters are updated adaptively by minimizing the CE between the estimated value and the expected value, until the residual converges and the switching precoding matrix is obtained. On this basis, the generalized power method and the least square method are used to give the phase shifts precoding matrix and the digital precoding matrix, respectively. Theoretical analysis and simulation results verify that the proposed method not only has lower computational complexity but also can provide better spectral efficiency and energy efficiency.