Reconfigurable intelligent surface (RIS)-aided tera-hertz (THz) communication is considered as a promising technique for the sixth-generation network. However, employment of hybrid beamforming in THz systems introduces beam split effect, resulting in severe achievable data rate loss. To realize the full potential of RIS-aided THz systems, it becomes crucial to acquire accurate channel state information. Therefore, in this paper, we propose a novel cyclic beam split generalized approximating message passing (CBS-GAMP) channel estimation scheme without requiring the knowledge of number of propagation paths. Specifically, we expand cascaded channels into sparse representations by designing a CBS dictionary, and then we propose the CBS-GAMP algorithm based on statistical inference framework. Numerical simulations demonstrate effectiveness of the proposed CBS-GAMP scheme against the existing solutions.
Generalized Approximating Message Passing Based Channel Estimation for RIS-Aided THz Communications with Beam Split
24.06.2024
463568 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch