Precoding method for minimum mean square error in multi-user multi-input multi-output system

The present invention discloses a multi-user MIMO downlink system based precoding method of the minimum mean square error, characterized in that the structure of the precoding matrix to the base station to all rules of the dummy overall channel matrix multiplied by the inverse of the user a block diagonal matrix is ​​calculated block diagonal matrix, using the obtained precoding matrix data stream as precoded and transmitted from the base station antenna based on a minimum mean square error criterion, each user itself received signal vector multiplying decoding a channel matrix into a plurality of single-input single output subchannels, each subchannel for data detection of the original data estimated. The present invention avoids the subscriber channels into a plurality of singular values ​​of subchannels SISO decomposition greatly reduces the computational complexity of precoding matrix without loss "and the rate of" performance, taking into account the system performance and algorithm implementation overhead.