Abstract: As a basis for precoder quantization, a channel matrix representing multi-antenna channel characteristics between a transmitter and a receiver in the mobile communications network is estimated (S1). A channel-assisted vector selection procedure (S2-S4) for iteratively selecting precoding vectors for a precoding matrix is performed. Each of a number of iterations of the novel iterative vector selection procedure generally involves evaluation (S2) of the performance of each one of a set of candidate vectors included in a given vector codebook with respect to a predetermined performance measure that is dependent on the estimated channel matrix, and selection (S3) of one of the set of candidate vectors as a respective precoding vector for the precoding matrix based on the evaluated performance. Information representative of a set of selected precoding vectors is then compiled (S5) to form a quantized representation of the precoding matrix for transmission to the transmitter side.

Abstract: As a basis for precoder quantization, a channel matrix representing multi-antenna channel characteristics between a transmitter and a receiver in the mobile communications network is estimated (S1). A channel-assisted vector selection procedure (S2-S4) for iteratively selecting precoding vectors for a precoding matrix is performed. Each of a number of iterations of the novel iterative vector selection procedure generally involves evaluation (S2) of the performance of each one of a set of candidate vectors included in a given vector codebook with respect to a predetermined performance measure that is dependent on the estimated channel matrix, and selection (S3) of one of the set of candidate vectors as a respective precoding vector for the precoding matrix based on the evaluated performance. Information representative of a set of selected precoding vectors is then compiled (S5) to form a quantized representation of the precoding matrix for transmission to the transmitter side.