| 摘要 |
The present invention provides a downlink channel estimation training sequence design method of frequency division duplex (FDD) large-scale multiple-input multiple-output (MIMO) system. In the system, a base station configured with large-scale multi-antenna array communicates with a single-antenna terminal user. The present invention performs continuous channel estimation utilizing spatio-temporal correlation between multi-antenna MIMO channel and on the basis of Kalman filtering. At each coherence time, the user performs Kalman filtering and prediction, obtains an estimated value of channel state information at the current coherence time, and gives a predicted value of the channel at the next coherence time. Statistical information of an intermediate variable is generated during the Kalman filtering and prediction, and the user feeds the necessary statistical information back to the base station via the uplink, so that the base station performs an optimal training sequence design. Simulation and analysis shows that, by the present solution, accuracy of channel estimation can be improved, while the time occupied by downlink training sequence of the system is reduced, thus improving the link throughout and system capacity. |
