Research On Acquisition Of Channel State Information In TDD Massive MIMO

Massive MIMO(Multiple-Input Multiple-Output) is one of the key potential technologies that enable the fifth-generation(5G) mobile communications systems. By deploying hundreds of antennas at the base station(BS), the massive MIMO system could achieve extremely high spectrum and energy efficiency. However, in massive MIMO systems, knowledge of channel state information(CSI) at BS is crucial to data detection in uplink and precoding in downlink. Therefore, the acquisition of CSI is a key issue for massive MIMO. Time division duplex(TDD) is an important scenario where the massive MIMO is studied. In this paper, we investigate the acquisition of CSI for uplink and downlink in TDD massive MIMO systems, separately.For the acquisition of CSI in uplink, we propose a novel pilot transmission scheme named staggered frame structure. Different from traditional synchronized frame structure, users send pilots successively under the staggered frame structure. What's more, two novel channel estimation algorithms are proposed for the staggered frame structure. Theoretical analysis and simulation results show that the massive MIMO system can achieve higher throughput with the proposed staggered frame structure and the proposed channel estimation methods.For the acquisition of CSI in downlink, we aims to restore the TDD channel reciprocity and then derive the downlink CSI from the uplink CSI. In massive MIMO system, the BS is deployed with hundreds of antennas. Due to the mismatches of receiving/transmitting radio frequency modules connected to the antennas, the TDD reciprocity is damaged in practical. We propose an assistant calibration device to calibrate the mismatches among the antennas, thus to restore the TDD reciprocity. Simulation results show that the downlink CSI could be obtained effectively by utilizing the assistant calibration device. Theoretical and simulation results have significant importance on the deployment of the assistant calibration device.