A Deep Learning Approach for Predicting Radio Channels Across Frequency in Mobile Networks

When operating massive multiple-input multiple-output (MIMO) systems with uplink (UL) and downlink (DL) channels at different frequencies (Frequency Division Duplex (FDD) operation), acquisition of Channel State Information (CSI) for downlink precoding is a major challenge. Since, barring transceiver impairments, both UL and DL CSI are determined by the physical environment surrounding the transmitter and receiver, it stands to reason that, for a static environment, a mapping from UL CSI to DL CSI precoder may exist. In this paper, we first, use generative-based Neural Networks to learn this mapping and provide baselines using signal processing approaches. Second, we use the trained models to test their generalization on unseen time, physical locations, and frequency gaps. All approaches are evaluated on the outdoor real-world dataset and under real-world scenarios.