Capacity analysis of MIMO systems with unknown channel state information

We consider a mobile wireless communication system composed of M transmit and N receive antennas operating in a fading environment. Assuming channel state information is unavailable to the transmitter and the receiver, a capacity upper bound of the unknown MIMO channel under the assumption of restricted input distributions is provided. By analyzing the proposed capacity upper bounds, we reinforce the advantages of using an orthogonal pilot structure which minimizes the mean square estimation error, in that it also maximizes the capacity upper bounds. Interestingly, the capacity upper bound is shown to be a monotonically decreasing function with respect to the number of pilot symbols, T_τ. Numerical evaluations of the capacity upper bound further demonstrate that the capacity gain is insignificant when T_τ decreases below M, suggesting an optimum training duration of M time slots.