Abstract
The present study considers a constant envelope (CE) precoding design problem in massive multiuser multiple-input multiple-output systems, where CE precoding is implemented with a network of finite resolution phase shifters (PSs). Mathematically, such CE precoding design problem can be formulated as a discrete optimization problem whose optimal solution can be obtained through an exhaustive search. Unfortunately, an exhaustive search involves exponential complexity with numerous base station antennas ( $M$ ) and bits in PS resolution. We resolve the aforementioned concern by using a low-complexity algorithm developed from the iterative gradient descent (IGD) based algorithmic framework. However, unlike conventional IGD-based solvers, finite resolution PSs are considered directly in the optimization procedure of the proposed algorithm to improve performance. In addition, the proposed algorithm is roughly $M/2$ times less complex than the conventional IGD algorithm at each iteration. The simulation results reveal that the proposed algorithm has nearly the same performance as that of the unquantized version of the state-of-the-art IGD algorithm.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have