Method and device for performing codebook processing on channel information

摘要

A method and a device for performing codebook processing on channel information are described, wherein channel information obtained by channel estimation is analysed to obtain a value space and a null space; the value space or the null space of which a dimension is small is taken as a codebook space; a codebook space vector is selected; and multidimensional codebook processing is performed separately on each codebook space vector, and information obtained by performing codebook is fed back to a base station. Thus the channel information feedback quantity and the codebook precision can be balanced flexibly without expanding codebook capacity, and the codebook precision is improved and channel information distortion is deduced under a premise that channel information data quantity is reduced.

主权项

1. A method for performing codebook processing on channel information, comprising:
analyzing the channel information obtained by channel estimation, to obtain a value space and a null space; taking the value space or the null space which has a small dimension as a codebook space; selecting one or more codebook space vectors; and performing multidimensional codebook processing separately on each of the one or more codebook space vectors, and feeding back information obtained by the codebook processing to a base station; wherein obtaining the value space and the null space comprises: performing singular value decomposition (svd) on the channel information obtained by the channel estimation, to obtain three matrixes which respectively are a left singular value matrix, a right singular value matrix, and a singular value matrix; and taking a vector of the right singular value matrix corresponding to a singular value in the three matrixes of which an absolute value is less than or equal to a preset first threshold as the null space, and taking a vector of the right singular value matrix corresponding to a singular value in the three matrixes of which an absolute value is larger than the preset first threshold as the value space; wherein taking the value space or the null space which has the small dimension as the codebook space comprises: judging whether vector quantity in the value space is less than or equal to vector quantity in the null space or not; if the vector quantity in the value space is less than or equal to the vector quantity in the null space, then taking the value space as the codebook space; and if the vector quantity in the value space is larger than the vector quantity in the null space, then taking the null space as the codebook space.