ADMM-Based Hybrid-Field Uplink Channel Estimation for Extremely Large-Scale MIMO Systems

Abstract

Extremely large-scale multiple-input multiple-output (XL-MIMO) is considered as a promising technique for upcoming sixth-generation (6G) mobile communication systems in improving the performance of current systems. However, XL-MIMO may face channel estimation challenges due to its high-dimensional channel. To reduce the pilot overhead for channel estimation, we exploit the channel sparsity in the uplink hybrid-field XL-MIMO system in 6G by utilizing both the far-field angle-domain dictionary and near-field polar-domain dictionary. Consequently, the channel estimation problem is formulated as an optimization problem with two ℓ0-norm constraints. Due to the non-convexity of the resulting problem, we relax the ℓ0-norm to ℓ1-norm and propose an alternating direction method of multiplier (ADMM)-based algorithm. Finally, simulation results demonstrate that the proposed ADMM-based method achieves more than 2 dB performance gain over existing schemes based on orthogonal matching pursuit.