Published in IET Communications Received on 26th May 2010 Revised on 18th July 2010 doi: 10.1049/iet-com.2010.0431 ISSN 1751-8628 Capacity of multiuser multiple input multiple output-orthogonal frequency division multiplexing systems with doubly correlated channels for various fading distributions S. Shyamala V. Bhaskar Department of Electronics and Communication Engineering, SRM University, Kattankulathur, Kancheepuram Dt. 603203, Tamil Nadu, India E-mail: vcharan@gmail.com Abstract: In this article, the capacity of multiuser multiple input multiple output (MIMO) orthogonal frequency division multiplexing (OFDM) system is analysed under Rayleigh, Rician, Lognormal and Weibull fading channels, and compared with each other under both semi- and doubly correlated channels. The distribution of the complex Wishart matrix is utilised to derive closed-form expressions of ergodic capacity of MIMO-OFDM systems under doubly correlated channels when the transmitter has no channel state information (CSI) and the receiver has perfect CSI. The capacity of MIMO-OFDM systems is plotted for various fading channels under different system specifications, such as number of transmit and receive antennas, number of subcarriers and different power levels. The obtained results are used to study the impact of channel correlation on capacity. 1 Introduction Wireless communication systems with multiple antennas at both the transmitter and the receiver have received considerable attention for the last decade owing the fact that multiple input multiple output (MIMO) systems can provide higher capacity as compared to single-input single-output (SISO) systems [1, 2]. The capacity of MIMO systems under flat fading in independent and identically distributed (i.i.d.) or correlated fading channels has been well examined in [1–3]. The capacity of MIMO systems under frequency- selective channels is studied in [4] in which orthogonal frequency division multiplexing (OFDM) is used to deal with frequency selective fading. In [5], MIMO capacity results of an OFDM-based system in correlated Rician fading has been presented. The ergodic capacity of spatially correlated Rician MIMO channels has been found in [6]. The case where the Rician component has a single dominant path and where the correlation occurs at one end of the MIMO link has been considered in [6]. In [7], the capacity of MIMO Rayleigh fading channels in the presence of spatial fading correlation at both the transmitter and receiver, assuming that the channel is unknown at the transmitter and perfectly known at the receiver, has been analysed. In [8], the eigenvalue densities of complex central Wishart matrices have been investigated. In particular, the capacity of MIMO Rayleigh distributed channels has been fully investigated. In [9], it has been shown that the expected log determinant of a complex non-central Wishart matrix is an increasing function of the non-centrality parameter which demonstrates that the mutual information (MI) corresponding to an isotropically distributed Gaussian input to a multi-antenna Rician fading channel is non-decreasing in the line-of-sight (LOS) component. The hypergeometric functions of the matrix arguments that occur in multivariate distributions, and also their expansions in zonal polynomials have been defined in [10]. Important properties of zonal polynomials and hypergeometric functions have been quoted. Formulas and methods of computing zonal polynomials up to degree 6 have been provided in [10]. In [11], new statistical properties of complex non-central Wishart matrices have been presented, which are used to derive bounds on the ergodic capacity of single-sided correlated Rician MIMO channels with arbitrary-rank channel mean matrices. In [12], MIMO channel capacity in correlated channel using exponential correlation matrix model has been investigated. It has been proved that an increase in correlation is equivalent to a decrease in the signal-to-noise ratio (SNR) for this model. In our work, the ergodic capacity for a multiuser MIMO- OFDM system is derived, plotted and analysed using the probability density function (PDF) of the complex central Wishart matrix. The paper is organised as follows: Section 2 discusses the system model. The capacity of multiuser MIMO-OFDM system with channel correlation is derived in Section 3. Numerical results are shown in Section 4. Finally, Section 5 presents the conclusions. Notations used in this paper are as follows: 1230 IET Commun., 2011, Vol. 5, Iss. 9, pp. 1230–1236 & The Institution of Engineering and Technology 2011 doi: 10.1049/iet-com.2010.0431 www.ietdl.org