Confidentiality analysis in multicast stochastic MIMO networks with imperfect CSIT

Application of multiple-input and multiple-output (MIMO) technology in wireless systems not only provides high spectral efficiency by the deployment of many number of antennas at both sides of a communication system for a given total transmit power, but it also enhances physical- layer security. Given the need for evaluation of realistic systems, this works studies the total information flow during multicast transmission in a region R, accounting for the unavoidable statistical spatial randomness of users locations, both in absence or presence of eavesdroppers under conditions of imperfect channel state information at the transmitter (CSIT) due to users' mobility. In particular, we first introduce a lower-bound of the space-time mutual information without any eavesdropping, and then we proceed with the derivation for both finite and infinitely large number of antennas. Moreover, in case that eavesdropping is present, we determine a metric, called space-time secrecy rate with imperfect CSIT, which characterizes the total amount of common confidential information flow per receive antenna, and we obtain it with asymptotic arguments in terms of the numbers of antennas at each side. Numerical results, validated by simulations, shed light on how the imperfect conditions affect the performance. Remarkably, imperfect CSIT may present benefits, since it engenders degradation in eavesdropper's performance.