Abstract: A communication device includes: a wireless transceiver configured to communicate with another communication device through a wireless channel; and at least one processor configured to execute instructions to: a) estimate at least one first channel response of the wireless channel based on a predetermined signal received from the another communication device; and b) predict a second channel response based on the at least one first channel response by an extrapolation method for signal transmission to the another communication device at a time instant where no channel estimation is performed.

Abstract: A base station, a method for a base station, and a program capable of appropriately selecting radio terminals for which spatial multiplexing is performed and an MCS, and thereby increasing throughput in multi-user MIMO are provided. A base station 11 according to the present disclosure includes: a correction parameter storage unit 1141 configured to store a plurality of correction parameters for each of a plurality of radio terminals 12, the plurality of correction parameters being provided for each of the radio terminals 12; a correction parameter selection unit 1131 configured to select a predetermined correction parameter from among the plurality of correction parameters based on a combination of radio terminals 12 that are spatially multiplexed in the same radio resource; a correction value calculation unit 1132 configured to calculate a correction value of the radio terminal 12 based on the predetermined correction parameter.

Abstract: A wireless receiving apparatus generates an N?×B? weight matrix W whose columns are orthogonal to each other by decomposing an estimated N?×M? channel matrix into the N?×B? weight matrix W and a B?×M? matrix containing two or more non-zero matrix elements in each column, where B? is an integer less than or equal to N?-1 and greater than or equal to M?. Alternatively, the wireless receiving apparatus generates a combination of an N?×B? sub-weight matrix W1 and at least one second sub-weight matrix, where the product of the sub-weight matrix W1 with the at least one second sub-weight matrix is equal to the weight matrix W. The wireless receiving apparatus performs receive beamforming on received signals of N? receiving antennas using the weight matrix W or the sub-weight matrix W1. The wireless receiving apparatus then performs a BP algorithm using receive-beamformed signals.

Abstract: A wireless receiving apparatus generates an N?×B? weight matrix W whose columns are orthogonal to each other by decomposing an estimated N?×M? channel matrix into the N?×B? weight matrix W and a B?×M? matrix containing two or more non-zero matrix elements in each column, where B? is an integer less than or equal to N?-1 and greater than or equal to M?. Alternatively, the wireless receiving apparatus generates a combination of an N?×B? sub-weight matrix W1 and at least one second sub-weight matrix, where the product of the sub-weight matrix W1 with the at least one second sub-weight matrix is equal to the weight matrix W. The wireless receiving apparatus performs receive beamforming on received signals of N? receiving antennas using the weight matrix W or the sub-weight matrix W1. The wireless receiving apparatus then performs a BP algorithm using receive-beamformed signals.

Abstract: A method implemented in a computer system includes training a network, which is obtained by unfolding an iterative algorithm for demodulation or demodulation and decoding, using a machine learning technique with a loss function that takes into account non-Gaussianity of a log-likelihood ratio (LLR) distribution calculated from an output of the network. The method further includes producing a first set of learned parameters of that iterative algorithm.

Abstract: A method implemented in a computer system includes training a network, which is obtained by unfolding an iterative algorithm for demodulation or demodulation and decoding, using a machine learning technique with a loss function that takes into account non-Gaussianity of a log-likelihood ratio (LLR) distribution calculated from an output of the network. The method further includes producing a first set of learned parameters of that iterative algorithm.