| Summary: | 碩士 === 國立成功大學 === 電腦與通信工程研究所 === 103 === Multiple-antennas and multiple-carriers systems have been attracted a plenty of attention for a long time due to the outstanding capacity, error performance and throughput. The transmitter can apply the precoding based on the fed-back channel state information (CSI). The beamforming vectors, which are essential CSI, could be applied in precoded multiple-input multiple-output (MIMO) systems. In order to efficiently feed back CSI, we propose some schemes that feed back the estimated channel responses (CRs) in the receiver to the transmitter based on the Karhunen-Loeve Expansion (KLE). The coefficients of the KLE are used as the parameters for rebuilding CRs. The number of coefficients of the KLE is much fewer than the size of CRs within a block, and this leads to efficient compression. These KLE coefficients are of Gaussian distribution. To efficiently feed back the coefficients for the MIMO system with orthogonal frequency-division multiplexing (OFDM), we use the Gaussian quantization (GQ) to reduce quantization error. We consider two types of one-dimensional blocks of CRs. We also apply the two-dimensional block of CRs. We develop efficient methods to estimate the multipath delays in the transmitter by using some previously fedback CRs. With the better estimated channel delays in the transmitter, one can reduce the delay of CR feedback and enhance its application on the precoding of the MIMO system.
|
|---|
