Analysis and Approximation of Statistical Distribution of Eigenvalues in英文文献.pdfVIP

IEICE TRANS. FUNDAMENTALS, VOL.E91–A, NO.10 OCTOBER 2008 2808 PAPER Special Section on Information Theory and Its Applications Analysis and Approximation of Statistical Distribution of Eigenvalues in i.i.d. MIMO Channels under Rayleigh Fading †a) †† † Tetsuki TANIGUCHI , Member, Shen SHA , Nonmember, and Yoshio KARASAWA , Fellow SUMMARY In multiple input multiple output (MIMO) communication In the ordered case, however, the derivation of marginal systems, eigenvalues of channel correlation matrices play an essential role distributions is not an easy problem. The closed form ex- for the performance analysis, and particularly the investigation about their pressions for the largest and the smallest eigenvalues are behavior under time-variant environment ruled by a certain statistics is an important problem. This paper first gives the theoretical expressions for the given in Refs. [5] to [10] (the largest) and [11] (the smallest), marginal distributions of all the ordered eigenvalues of MIMO correlation respectively. The eigenvalue distribution in MIMO chan- matrices under i.i.d. (independent and identically distributed) Rayleigh fad- nel is studied also for correlated case [12], but the explicit ing environment. Then, an approximation method of those marginal distri- forms of marginal distributions of other eigenvalues are not butions is presented: We show that the theory of SIMO space diversity us- found in the literature even in the uncorrelated case. More- ing maximal ratio combining (MRC) is applicable to the approxi