Modified Geometric Mean Decomposition and Orthogonal Matching Pursuit Based Hybrid Precoding for Millimeter-Wave Massive MIMO Systems

Abstract

The hybrid precoding technique efficiently achieves the reduction of the number of chains in the radio frequency range namely the RF chains in mm-wave massive MIMO system. The method of combined hybrid precoding that is existent now is based on Singular Value Decomposition algorithm which gives rise to the problem of complication in the bit allocation in order to match the varying signal-to-noise ratio for each sub-channel. Thus, this problem of complexity in bit allocation could be easily sorted out with the proposed method of modified Geometric Mean Decomposition and orthogonal matching pursuit algorithm. This is achieved by manipulating and bringing up an identical signal-to-noise ratio enabling an easy and simplified bit allocation. But when the system is completely digital, the problem of increased number of RF chains sprouts out. Thus, the hybrid precoder which is a combination of analog and digital precoder comes into the picture. This pair of analog and digital precoder sufficiently brings out similar output as that of the unconstrained6fully8digital precoder. In order to achieve this initially, the analog precoder is designed and then fixed. This is then followed by the designing of digital precoder. Orthogonal Matching Pursuit algorithm is fetched to design the analog precoder while GMD is used for the design of digital precoder. The results after simulation of this algorithm are highly efficient and achieve a remarkably better performance than the existing methods.