A fast gridless mmWave full-dimensional MIMO channel estimation method

Abstract

Most existing channel estimation methods for millimeter-wave (mmWave) multiple-input multiple-output (MIMO) communication systems focus on linear arrays rather than planar arrays since discretizing the high-dimensional angle space in planar array scenario brings in high computations. In this paper, we consider the channel estimation problem for mmWave full-dimensional MIMO system where the base station (BS) and user equipment (UE) are equipped with planar arrays. We first propose an atomic ℓ0 norm based optimization problem to find the angle estimates and then transfer the problem into a solvable semidefinite programming (SDP). We next propose a computationally efficient algorithm to solve the SDP rather than employing existing standard solvers. Compared with the grid based methods which require discretizing the angle space, our method avoids angle discretization which brings in huge computations. Various simulations show that our method is able to provide satisfying channel estimation performance while is also computationally efficient.