认知星地网络中鲁棒多目标优化安全波束形成算法

Robust multi-objective beamforming algorithm in cognitive satellite-terrestrial network

在认知星地网络中,对于系统多个目标之间的物理层安全问题,提出了基于鲁棒多目标的安全波束成形算法。首先,在满足用户最小可达安全速率约束和基站最大发射功率限制的约束下,研究了基站总发射功率最小化和次级用户可达安全速率最大化的联合优化问题。然后,考虑到不完全的信道状态信息的实际假设,提出了基于加权切比雪夫方法的多目标鲁棒波束成形方案。其次,通过利用半正定规划和泰勒级数展开等方法将非凸的优化问题转化为具有线性矩阵不等式的线性问题,来获得最优权向量和人工噪声向量,从而得到了两个冲突问题之间的资源分配策略的帕累托最优集。最后,仿真实验的数值结果验证了所提出算法的有效性和优越性。

This paper investigates a robust multi-objective optimization(MOO) beamforming in the cognitive satellite-terrestrial network,where the eavesdroppers attempts to intercept the private information intended for terrestrial user.Assuming that imperfect channel state informations are known at base station,we adopt the weighted Tchebycheff approach to optimize the multi-object optimization problem between achievable secrecy rate maximization and transmit power minimization.Then,by introducing several auxiliary variables and applying semidefinite programming and means of Taylor approximation,the original non-convex MOO problem is transformed into one with a series of linear constraints to obtain complete Pareto optimality set.Finally,simulation results illustrate the tradeoff between the conflicting objectives and the superiority of the proposed algorithm compared with other schemes.