Probabilistic-based secrecy rate maximisation for MIMOME wiretap channels: Towards novel convexification procedures

The paper first provides a general overview on the existing probabilistic-driven methods aimed at convexifying the secrecy rate maximisation problem. Challenges and requirements of the procedures are fully taken into consideration for multiinput multioutput multiantenna eavesdropper wiretap channels. The problem constrained by an outage probability for the pure Ergodic capacity (perfect secrecy defined by Shannon) is analytically solved. Such analytical solution is performed robustly, something that originates principally from our imperfect statistical knowledge of the wiretap channel at the transmitter. Then, solely concentrating on McDiarmid inequality, we also extend our previous work to propose 2 new approaches. The bounds of our represented schemes are obviously accurate, tight, and achievable, something that effectively guarantee the system to overcome the probable challenges, eg, correlation in multiinput multioutput antenna arrays. Furthermore, the procedures are straightforwardly compared with each other from standpoint of complexity. Results are technically conducted to assure us about the proposed methods' merits, merely defining the significantly more favourable asymptotic bounds and the less resultant complexities. In particular, our proposed approaches sensibly outperform all the procedures, per se, something that can be accordingly assigned for other ergodic equilibria.