Secure Relay Beamforming Algorithm Research In Wireless Relay Networks

In this thesis,we mainly research the problem of secure relay beamforming in wireless relay networks.In the whole thesis,we assume that the channel state information is perfect.The whole mostly consists of two parts and details as follow.the first part studies the problem of secure relay beamforming in two-way relay cognitive radio networks which applies the technology of simultaneous wireless information and power transfer.Our destination is,under the conditions of the transmission power constraints of the relay and the constraints of the energy harvesting and the interference constraints of primary user,to maximize the average secure rate of the communication system.However,due to the objection function is non-convex,it is hard that employing the methods of convex optimization to solve our problem.Because of this reason,we propose two solutions to overcome above difficulty.First,we propose a solution that base on rank-one relaxation two dimension to get the upper bound of the communication system.In order to reduce computational complexity,we present another solution base on sequence parameters convex approximation iteration to get the local Solution of communication system.Simulation results demonstrates that base on sequence parameters convex approximation iteration can get a closer performance than base on rank-one relaxation two dimension,and it has lower computational complexity.The second part,we investigate joint power allocation and secure relay beamforing problem in a non-orthogonal multiple access amplify and forward relay networks.We try our best to maximize our objective problem which is maximize achievable secrecy rate of the user which should has the best channel condition.Our problem constrained by the communication system achievable rate and some individual maximize transmit power.Since the optimization problem is non-convex,we propose a method which employs one-dimensional search iteration algorithm based Rank-one relaxation.In order to reduce computational complexity,we transform the problem into a sequential parametric convex approximation based iterative algorithm to find a locally optimal solution.Furthermore,we also proposed a second-order cone programming based iterative algorithm.Simulation results demonstrate that the SOCP based iterative algorithm,which has lower computational complexity than one-dimensional iterative algorithm,has the performance close one-dimensional iterative algorithm.The SPCA based iterative algorithm has lowest computational complexity among proposed methods,however,it's performance is worse.