Abstract:
This paper investigates the recovery of a spectrally sparse signal (SSS) from partially observed entries, with particular emphasis on computational efficiency for large s...Show MoreMetadata
Abstract:
This paper investigates the recovery of a spectrally sparse signal (SSS) from partially observed entries, with particular emphasis on computational efficiency for large scaled problems. We formulate the SSS recovery as a nonconvex low-rank Hankel matrix recovery problem. A projected proximal gradient method has been developed. It is an iterative process where each iteration involves two steps. The first subspace projection step finds the optimal solution in a low-rank and Hankel matrix space for given column and row subspaces. The second step optimizes all involved variables including the column and row subspaces by using a proximal gradient process. In both steps, sub-problems are formulated so that both the low-rank and the Hankel structures are fully exploited for computational efficiency. The combination of these two steps substantially improves the convergence rate. Numerical simulations demonstrate a significant improvement in efficiency compared with the benchmark algorithms.
Date of Conference: 04-08 September 2023
Date Added to IEEE Xplore: 01 November 2023
ISBN Information: