CaFTRA: Frequency-Domain Correlation-Aware Feedback-Free MIMO Transmission and Resource Allocation for 6G and Beyond

The fundamental design of wireless systems toward AI-native 6 G and beyond are driven by the need for ever-increasing demand of mobile data traffic, extreme spectral efficiency and adaptability across diverse service scenarios. To overcome the limitations posed by feedback-based multiple-input and multiple-output (MIMO) transmission, we propose a novel frequency-domain Correlation-aware Feedback-free MIMO Transmission and Resource Allocation (CaFTRA) framework tailored for fully-decoupled radio access networks (FD-RAN) to meet the emerging requirements of AI-Native 6 G and beyond. By leveraging artificial intelligence (AI), CaFTRA eliminates the need for real-time channel state information (CSI) feedback during online MIMO operation by predicting CSI from user geolocation. We introduce a learnable-queries-driven Transformer network for geolocation-to-CSI mapping, which utilizes multi-head attention and learnable query embeddings to accurately capture frequency-domain correlations among resource blocks (RBs), thereby significantly improving CSI prediction precision. Once base stations (BSs) adopt feedback-free transmission, their downlink transmission coverage can be significantly expanded due to the elimination of frequent uplink feedback. To fully explore the possibility of such extensive coverage with multi-BS cooperation, we apply a low-complexity many-to-one matching theory-based algorithm, which is proved to converge to a stable matching within limited iterations. Simulation results based on the Vienna 5 G system-level simulator show that CaFTRA achieves stable matching convergence and significant gains in spectral efficiency and user fairness over representative 5 G baselines, while remaining effective across different deployment scenario and carrier band through scenario-specific retraining.