A Novel Codebook Design for Linear Precoding Systems

摘要

Background Precoding has been regarded as an effective method to improve the performance of multiple-input multiple-output (MIMO) systems. Many works have been investigated for designing the codebooks of precoding matrices, where the discrete Fourier transform (DFT) codebook have been proposed as the optimal one. However, it can be specified that the construction of DFT codebook leads to a high complexity. Hence, finding a different codebook, which needs a lower construction complexity and less performance loss, is the main purpose of this paper. Materials and Methods It is found that the orthogonal space time block codes (OSTBC) provides many orthogonal signal matrices, which almost satisfy property of precoding codebooks. Therefore, authors will evaluate the complexity and performance by using OSTBC codebooks. More specifically, the receiver chooses a matrix from the OSTBC codebook and feedbacks the optimal codebook matrix to the transmitter. Then, the authors compare the complexity with the conventional discrete Fourier transform (DFT) codebook. Results From the designed equations of OSTBC, it can be seen that the construction of OSTBC codebook is simpler. Simulation results show that using the proposed OSTBC codebook leads to the almost same performance compared with that of the DFT codebook. Conclusions The proposed OSTBC codebook is advisable to be considered as a tradeoff between the complexity and performance. Main conclusions and interpretation of findings In this paper, the authors propose a novel codebook design for linear precoding coding scheme. In the proposed scheme, it specifies a limited feedback method that uses orthogonal space time block codes (OSTBC) as the codebook of precoding matrices, which is known to both the transmitter and receiver. Compared with the conventional discrete Fourier transform (DFT) codebook, OSTBC codebook require less complexity and achieves the nearly same performance. Note that the proposed method fails to obtain a better performance compared with the DFT codebook, however, the simple implementation is still an attractive advantage.