Affine Precoding and Decoding MIMO Frequency-selective Fading Channel for Estimation and Source Detection

A new affine precoding and decoding method for multiple-input multiple-output (MIMO) frequency-selective fading channel is proposed. Based on a concept of orthogonal superimposed training on linear precoding, the optimal design of an affine precoder consisting of linear precoder and training sequence is developed. The training sequence is superimposed on the linearly precoded data, i.e., it is arithmetically added to the precoded data. By additionally employing decoding, the channel can be estimated independently of the precoded data without consuming transmission bandwidth for the training sequence. Furthermore, a power allocation scheme is designed to determine the optimal ratio of power between the training sequence and the data. It is shown through simulation that the proposed method has considerable advantages over other affine precoding methods with regard to source detection improvement and reduced computational complexity