Secrecy Capacity of Artificial Noise Aided Secure Communication in MIMO Rician Channels

Recent research on the physical layer security of wireless systems focuses on artificial noise (AN) aided security. The main metric for analysis of such systems is the secrecy capacity of the system. Most of the AN schemes proposed in recent research are based on a hypothesis that the number of transmit antennas is larger than that of the receive antennas. Under this assumption, the system can utilize all eigensubchannels, equal to the number of the receive antennas, to send secret messages. The remaining null spaces are used for transmitting AN signals. These AN signals null out at the legitimate receivers and degrade illegitimate receiver's channels. However, this strategy can significantly impair the secrecy capacity of the system if the number of transmit antennas is constrained or even smaller than the number of receive antennas. Recently, a new strategy has been proposed, where messages are encoded in s (which is a variable) strongest eigen-subchannels based on ordered eigenvalues of Wishart matrices, while AN signals are generated in remaining spaces. This paper extends this strategy to Rician channels using the complex non-central Wishart distribution. A closed form expression for secrecy capacity of such system is computed using majorization theory. Performance of a MIMO communication system in Rician fading environment is simulated and effect of the Rician factor on secrecy capacity is studied. The same approach is then extended to decode and forward relay network. Secrecy capacity of the said network is computed, and the effect of AN is studied using the same methodology.