Low-PAPR Precoding Design for Massive Multiuser MIMO Systems via Riemannian Manifold Optimization

Constant envelope (CE) precoding is of great interest in the context of massive multiple-input multiple-output antenna systems, because it can satisfy the demand of hardware-friendly implementation. However, existing solvers either fall into a local minimum or require huge computational cost because of the non-convex nature of the CE precoding design problem. Some computationally efficient methods are available that can relax the CE constraints to achieve the desired performance for suppressing multiuser interference (MUI) but at the cost of reducing the peak-to-average power ratio performance. To improve the MUI performance and to consider computational complexity, we propose the use of Riemannian manifold optimization (RMO) that can solve the CE precoding design problem. The simulation results demonstrate that the proposed RMO-based solver achieves nearly the same performance of the conventional sequential gradient descent-based solver, but the complexity is considerably reduced.