Super-Resolution Channel Estimation for Movable Antennas via Atomic Norm Minimization

Movable antenna (MA) technology has recently emerged as a promising solution for enhancing wireless communication performance through spatial adaptability and reconfigurability. To fully leverage this potential, it is essential to obtain accurate channel state information (CSI) across the antenna’s movable region, which is practically difficult to address due to its extremely high pilot overhead. In this paper, we investigate the channel estimation problem in an MA system and propose a two-stage estimation framework based on atomic norm minimization. Unlike conventional on-grid compressed sensing (CS) methods, our proposed approach operates based on an infinite atom set, enabling super-resolution estimation of the channel parameters without grid mismatch errors. Specifically, in the first stage, the angles of departure (AoDs) are estimated using atomic norm minimization from pilot measurements at different antenna positions. In the second stage, the path gains are estimated using the least squares method, with the antenna positions optimized to minimize the mean square error. Numerical results demonstrate the superior accuracy of the proposed framework in estimating both AoDs and overall channel responses, significantly outperforming traditional on-grid CS benchmarks.