Hybrid precoding with compressive sensing based limited feedback in massive MIMO systems

Dongmei Jiang,Balasubramaniam Natarajan
DOI: https://doi.org/10.1002/ett.3108
IF: 3.6
2016-09-16
Transactions on Emerging Telecommunications Technologies
Abstract:Massive multiple‐input multiple‐output (MIMO) systems are an enabling technology for high capacity 5G cellular networks. However, the high hardware complexity and power consumption make it difficult to apply conventional MIMO precoding schemes in massive MIMO. Additionally, thanks to the large channel dimension, applying traditional limited feedback schemes for obtaining channel state information at the transmitter is a challenge. In this paper, we develop a low‐complexity hybrid precoding scheme for multiuser massive MIMO system in mmWave communication. We design the analog beamformer as the conjugate transpose of the downlink channel from base station to the users. Then zero‐forcing baseband processing is applied over the effective channel. We get the optimal power allocation policy and compare its performance with average power allocation based on channel state information at transmitter. The channel state information at transmitter is obtained via a compressive sensing based limited feedback strategy. The effectiveness of hybrid precoding relies on the feedback quality of channel state information, which relies on the compression ratio. We quantify the impact of compression ratio on precoder performance and derive an upper bound on the sum rate. Simulation results illustrate the performance of our proposed scheme and the effect of compressive sensing based limited feedback. This paper develops a low complexity hybrid precoding scheme for multi‐user massive MIMO system in mmWave communication. The effectiveness of hybrid precoding relies on the compressive sensing‐based feedback quality of channel state information, which relies on the compression ratio. By derivations and simulation results, we show the performance of our proposed schemes.
telecommunications
What problem does this paper attempt to address?