Near-Field Channel Estimation and Localization with Coprime Arrays

This paper addresses uplink near-field, semi-blind, channel estimation and localization for Extremely Large Coprime Antenna Arrays explicitly accounting for multipath propagation. We develop a subspace-based estimation strategy that jointly exploits pilot and data symbols to recover channel responses and localization parameters. Our approach is compared to the conventional Uniform Linear Array (ULA). While ULAs provide slightly better accuracy under ideal conditions with the same aperture, Coprime Arrays (CAs) require significantly fewer elements, reducing computational complexity, and are more robust to mutual coupling (MC). Overall, CAs provide a practical and efficient solution for near-field channel estimation and localization for 6G communication systems.