The purpose of this paper is to examine the possibilities of increasing the channel capacity in tunnels, and particularly in railway tunnels, due to the use of multiple-input-multiple-output techniques. Many measurement campaigns have been carried out, considering the complicated geometric structure of these tunnels. We determine the channel characteristics by making a statistical analysis of correlation matrices between antennas and singular values of the channel transfer matrix. A comparison between the theoretical and experimental results is also presented. A stochastic channel model is then proposed and validated.

Both a theoretical approach and measurements of channel characteristics have been carried out to determine the maximum capacity which can be expected when MIMO techniques are implemented in tunnels, and especially in subway tunnels. Emphasis is given to the influence of the position of the antennas inside the transmitting and receiving arrays on the correlation between paths and on the singular values of the transfer matrix. The channel capacity is given for various geometrical configurations and compared with the case of i.i.d. Rayleigh channels.

Extensive measurements have been carried out in Paris subway tunnels to determine the channel characteristics and to deduce the available channel capacity of a multiple input multiple output (MIMO) transmission system. An analysis of the power correlation coefficients of the antennas on the train and on the platform is also performed to optimize the position of the antennas.