Enabling Feedback-Free MIMO Transmission for FD-RAN: A Data-driven Approach

• DOI: 10.48550/arxiv.2311.14329
• Publication Date: 2023-01-01
• Authors: Jingbo Liu; Jiacheng Chen; Zongxi Liu; Haibo Zhou
• Keywords: Channel state information

To enhance flexibility and facilitate resource cooperation, a novel fully-decoupled radio access network (FD-RAN) architecture is proposed for 6G. However, the decoupling of uplink (UL) downlink (DL) in FD-RAN makes existing feedback mechanism ineffective. this end, we propose an end-to-end data-driven MIMO solution without conventional channel procedure. Data-driven can alleviate drawbacks including overheads delay, provide customized precoding design different BSs based on their historical data. It essentially learns mapping from geolocation to transmission parameters. We first present codebook-based approach, which selects parameters statistics discrete state information (CSI) values utilizes integer interpolation spatial inference. further non-codebook-based 1) derives optimal precoder singular value decomposition (SVD) channel; 2) variational autoencoder (VAE) select representative latent Gaussian representations; 3) exploits process regression (GPR) predict unknown precoders space domain. Extensive simulations are performed link-level 5G simulator using realistic ray-tracing The results demonstrate effectiveness MIMO, showcasing its potential application