Secrecy Capacity Of Wiretap Channel For MIMO Cognitive Radio

As an intelligent wireless technology,cognitive radio can effectively improve the spectral efficiency,and by combining cognitive radio with multiple-input multiple-output(MIMO)technology,the degree of freedom for cognitive user's channel matrix can be significantly increased so that the channel capacity can be increased.However,due to the random and broadcast nature of wireless channels,cognitive radio system can be easily eavesdropped by malicious users,thus giving rise to security issues.In recent years,physical layer security has attracted wide attention in wireless communication and information security.Compared with traditional encryption method,physical layer security has lower complexity and stronger feasibility so that the legitimate user can effectively avoid or decrease the threaten of being eavesdropped.In this dissertation,the key technology of physical layer security in MIMO cognitive radio has been deeply researched,the main research work and the contribution are summarized as follows:1.The mathematical MIMO cognitive wiretap channel model has been built,and on the condition of full channel state information at cognitive transmitter,the secrecy capacity theorem of this wiretap channel model has been established.In this theorem,the secrecy capacity can be expressed as a non-convex max problem as well as a convex-concave max-min problem.And the secrecy capacity is derived and proved according to information theory,the upper and lower bound of secrecy capacity have been derived,the optimal distribution of input signal has been defined,the saddle point as the optimal solution of max-min problem and the upper bound coincides with lower bound under saddle point have been proved.Based on these results,the secrecy capacity of MIMO cognitive radio wiretap channel with multiple PRs and multiple evesdroppers has been studied.Simulation results have verified the correctness of the secrecy capacity theorem.2.The properties of secrecy capacity of MIMO cognitive radio wiretap have been studied.The relative properties of optimal transmit covariance have been derived,including the optimal signaling space,the upper bound of the rank of the optimal covariance,and the sufficient condition of optimal beamforming.The sufficient and necessary conditions of non-zero and unbounded growth of secrecy capacity have been derived.The interplay between the status of total power constraint(TPC)as well as interference power constraint(IPC)and transmit power threshold as well as interference power threshold have been analyzed.Simulation results have verified the correctness of the upper bound of the rank of the optimal covariance,the sufficient and necessary conditions of unbounded growth of secrecy capacity,and also verified the correctness of analysis of interplay between the statues of TPC and IPC and the corresponding thresholds.3.Based on the modeling of strictly degraded MIMO cognitive radio wiretap channel,the numerical solution of secrecy capacity has been derived in an analytical way.Based on non-zero interference power threshold setting,a method of full rank solution has been derived as well as the sufficient conditions of optimality of full rank;Based on zero interference power threshold setting,a method of rank deficient solution has been derived,and the optimal signaling space,the sufficient conditions of optimal beamforming and the unbounded growth of secrecy capacity have also been derived.Simulation results have been verified the correctness of the proposed full rank and rank deficient solutions,they have same performance with Monte Carlo method.4.The MIMO cognitive radio wiretap channel model under the condition of multiple primary receivers and multiple eavesdroppers has been built,and two algorithms with global convergence for computing optimal input transmit covariance has been proposed to solve the secrecy capacity optimization problem.For the case of all non-singular primary user channel matrices,an algorithm is proposed based on barrier method,and in combination with Newton method and backtracking line search method.For the case of existing singular PU channel matrices,an algorithm is proposed based on bisection search method.Mathematical Analysis and derivation have verified the global convergence of the two algorithms.Simulation results show that both algorithm have high accuracy and fast convergence,and they have significantly higher performance than the alternating optimization method with only stationary convergence.