MIMO precoding performance for correlated and estimated Rician fading

Abstract

In recent years, there have been significant breakthroughs in wireless communication systems design based on the multiple-input multiple-output (MIMO) concept. Exploiting the channel state information at the transmitter (CSIT) can improve the link performance. Thus, MIMO precoding has been an active research area for wireless communications. By specifying a precoding matrix with full CSIT, one can activate the strongest channel modes. In practice, important issues include how to obtain an accurate estimated CSI and use it. In this paper, we exploit estimated CSI for a correlated realistic Rician fading channel with estimated statistics. Then, a new precoder by providing the statistical mean of the channel matrix is presented. Numerical results show the performance degradation exists between the typical channel assumption, i.e., Rayleigh fading, and the realistic channel, i.e., correlated Rician fading. The new CSIT by sufficiently considering the channel statistics can reduce the error rate. The rank of the Rician fading channel matrix renders performance sensitive for the precoding performance.