Random coupling model for wireless communication channels

In this paper, we derive an explicit formula for the voltage-to-voltage transfer function of multiple-input multiple-output (MIMO) wireless channels. A statistical model, the random coupling model, is used to develop the open-circuit transfer function of the MIMO channel on a physical basis. The emulation of realistic wireless channels is typically performed through irregular cavities with high losses. In this case, we find that the transfer function takes a simple form involving the free-space impedance matrix of antennas and a fluctuation matrix expressing the wave chaos inside the environment. Monte Carlo simulations of the open-circuit transfer function are performed for MIMO systems up to three antennas in the transmit and receive arrays. In contrast to the common assumption that the MIMO channel fulfill multivariate normality (MVN), the Hans-Zinckler test of the obtained ensembles of the MIMO channel show that the MVN assumption of the MIMO channel tends to be invalid with an increasing number of antennas in the transmitting and receiving arrays, when mutual coupling is present in the arrays. Numerical results indicate that this effect is more pronounced at relatively low frequencies.