Geospatial Multi-User Channel Condition Interpolation for Instantaneous Sporadic Transmission and CSI Feedback Overhead Reduction in 5G and Beyond

Autor: Jasna Jankovic, Zeljko Ilic, Gordan Sisul
Jazyk: angličtina
Rok vydání: 2024
Předmět:
Zdroj: IEEE Access, Vol 12, Pp 186883-186897 (2024)
Druh dokumentu: article
ISSN: 2169-3536
DOI: 10.1109/ACCESS.2024.3515286
Popis: Next-generation networks look for innovative solutions to enable new service functionalities while increasing the system efficiency. This paper proposes an innovative approach for the instantaneous transmission of sporadic user traffic and Channel State Information (CSI) feedback overhead reduction based on the geospatial multi-user channel condition interpolation at the Base Station (BS). The approach relies on the fact that unknown channel conditions of some users can be estimated (i.e., interpolated) at the BS based on known channel condition values of other users, thus omitting the need for them to send the CSI feedback. We conduct a practical analysis of the applicability and performance of the proposed approach for one user in different propagation environments and using different interpolation methods. Further, we elaborate our findings on the multi-user scenario by considering the overall system performance with respect to the ratio of users whose channel conditions are being estimated. The simulation results show the best performance of the proposed approach in high-user-density and low-channel-variability environments, where the best interpolation precision is achieved. Instantaneous transmissions obtain transmission reliability between 68% and 100%, depending on user density and propagation environment, with the best performance achieved in the suburban environment. CSI feedback overhead could be reduced significantly at the cost of acceptably low downlink performance degradation (e.g., 3% downlink performance degradation for a 10% overhead reduction), especially in the dense urban environment where uplink gets saturated with the CSI feedback.
Databáze: Directory of Open Access Journals