With the aim of boosting the security of the conventional directional modulation (DM) network, a secure DM network assisted by intelligent reflecting surface (IRS) is investigated in this paper. To maximize the security rate (SR), we jointly optimize the power allocation (PA) factor, confidential message (CM) beamforming, artificial noise (AN) beamforming, and IRS reflected beamforming. To tackle the formulated problem, a maximizing SR with high-performance (Max-SR-HP) scheme is proposed, where the PA factor, CM beamforming, AN beamforming, and IRS phase shift matrix are derived by the derivative operation, generalized Rayleigh-Rize, generalized power iteration, and semidefinite relaxation criteria, respectively. Given that the high complexity of the above scheme, a maximizing SR with low-complexity (Max-SR-LC) scheme is proposed, which employs the generalized leakage and successive convex approximation algorithms to derive the variables. Simulation results show that both the proposed schemes can significantly boost the SR performance, and are better than the equal PA, no IRS and random phase shift IRS schemes.

翻译：为提升传统定向调制网络的安全性，本文研究了一种智能反射面（IRS）辅助的安全定向调制网络。为最大化安全速率（SR），我们联合优化了功率分配（PA）因子、机密消息（CM）波束赋形、人工噪声（AN）波束赋形以及IRS反射波束赋形。针对所构建的问题，提出了一种高性能最大化安全速率（Max-SR-HP）方案，其中PA因子、CM波束赋形、AN波束赋形及IRS相位偏移矩阵分别通过求导运算、广义Rayleigh-Rize方法、广义幂迭代法和半定松弛准则求解。鉴于上述方案复杂度较高，进一步提出了一种低复杂度最大化安全速率（Max-SR-LC）方案，该方案采用广义泄漏和逐次凸逼近算法求解变量。仿真结果表明，所提两种方案均能显著提升安全速率性能，且优于等功率分配、无IRS及随机相位偏移IRS方案。