Nokia proposes new method for channel estimation in 5G networks
Nokia Bell Labs has published research on a new method to improve channel estimation in FDD Massive MIMO systems for 5G networks with low reference signal overhead.
Nokia Oyj's research arm, Nokia Bell Labs, has presented a new technique aimed at enhancing channel estimation within Frequency Division Duplex (FDD) Massive MIMO systems for 5G networks. The initiative seeks to significantly boost data throughput in upcoming 5G deployments, particularly in the sub-6 GHz spectrum where FDD is expected to remain prevalent.
The proposed solution, detailed in the publication "Low-Overhead Cyclic Reference Signals for Channel Estimation in FDD Massive MIMO," addresses the need for efficient channel estimation in systems with a large number of antennas. It consists of three key components: leveraging a fixed grid of beams to focus estimation on relevant signals, employing cyclic reference signal sequences that repeat over time, and enabling terminals to estimate only the most pertinent channels for their connection.
Bell Labs investigated both linear mean square estimation and Kalman estimation techniques. These methods utilize frequency and antenna correlations, with the Kalman approach also incorporating temporal correlation. Extensive simulations indicate that this scheme yields channel estimates leading to negligible degradation in beamforming performance compared to having complete channel knowledge, while minimizing the overhead associated with reference signals.