Deep Learning for Massive MIMO CSI Feedback
share
In frequency division duplex mode, the downlink channel state information (CSI) should be conveyed to the base station through feedback links so that the potential gains of a massive multiple-input multiple-output can be exhibited. However, the excessive feedback overhead remains a bottleneck in this regime. In this letter, we use beep learning technology to develop CsiNet, a novel CSI sensing and recovery network that learns to effectively use channel structure from training samples. In particular, CsiNet learns a transformation from CSI to a near-optimal number of representations (codewords) and an inverse transformation from codewords to CSI. Experiments demonstrate that CsiNet can recover CSI with significantly improved reconstruction quality compared with existing compressive sensing (CS)-based methods. Even at excessively low compression regions where CS-based methods cannot work, CsiNet retains effective beamforming gain.
READ FULL TEXT