Capacity analysis of MIMO rayleigh channel with spatial fading correlations and estimation errors

Abstract

Multiple Input Multiple Output (MIMO) communications systems equipped with antennas array at both the transmitter and receiver ends are promising schemes to realize higher rate and/or reliable data transmitter and receiver ends are promising schemes to realize higher rate and/or reliable data transmission. In this thesis, capacity analysis of MIMO Rayleigh Channel with spatal correlation at the receiver of multipath is studied. In general, a model configuration of local scatterng around a mobile station in MIMO environment is carried out by simulaton to examine spatial correlation coefficients. Based on statistical properties of the eigenvalues of the eigenvalues of correlated complex random Wishart matrices, the exact closed-form expressions of distribution of the eigenvalues are investigated. Thaen, the eneral closed-form evaluation of integral form is proposed based on Meijer's G-function. The analytical results demonstrate that the ergodic capacties are improved by increasing the number of the antennas and signal-to-noise ratio (SNR's). Compared with independent identically distributed (iid) Rayleigh channel, the incremental improvement of correlated Rayleigh channel is reduced by spatial fading correlation. The analytical results validated by Monte-Carlo simulations that show a good agreement. In addition to the capacity analysis, the aspect of information theory of a practical MIMO transmission scheme is discussed, namely, spatial multiplexing MIMO system with zero forcing receiver which operating under realistic conditions such as spatially correlation, channel estimation errors, and delay feedback from the receiver to the transmitter. Results are presented for spatially correlated Rayleigh fading channels and estimation errors. It is demonstrated that the capacity of the system suffers degradation when the channel state information (CSI) is not perfect and in spatially correlated signals. A lower bound on mutual information is also derived for flat fading channel with the effective noise power matrix, which lead to simple study of insights.