Secure Transmission Scheme for Reconfigurable Intelligent Surface-enabled Cooperative Simultaneous Wireless Information and Power Transfer Non-Orthogonal Multiple Access System
-
摘要: 可重构智能超表面(RIS)因能提供额外的无源波束增益被视为一项颇具前景的技术。考虑到未来大型物联网中不同用户服务需求的多样性及信息传输的安全性,该文面向协作无线携能同传-非正交多址接入(SWIPT-NOMA)系统,提出一种RIS使能的安全传输方案。通过合理部署RIS的位置,将其同时作用于直接传输阶段和协作传输阶段。在满足非正交多址接入(NOMA)弱用户信息速率需求、NOMA强用户能量收集需求和基站最小发射功率的条件下,通过联合优化基站的有源波束成形、RIS的相移矩阵、强用户的功率分割系数等来最大化强用户的保密速率。为解决所提的多变量耦合的非凸优化问题,该文基于交替迭代优化算法,对基站的有源波束成形、直接传输阶段的RIS无源波束相移矩阵、协作传输阶段的RIS有源波束相移矩阵以及强用户的功率分割系数等进行了多次交替迭代优化,直至算法收敛。仿真结果验证了该文算法的收敛性,且与其它基准方案相比,所提方案可进一步提高强用户的保密速率。Abstract:
Objective The Reconfigurable Intelligent Surface (RIS) is emerging as a promising technology due to its ability to provide passive beamforming gains, which can be seamlessly integrated into existing wireless networks without altering physical layer standards. The integration of RIS with other advanced technologies offers new opportunities for communication network design. In the context of future large-scale Internet of Things (IoT) systems, users are expected to have diverse requirements. These differences in structure and function lead to two distinct receiver operation modes: Power Splitting (PS) and Time Switching (TS). Furthermore, users’ service needs may vary, including energy harvesting and information transmission. In practice, IoT terminals often face energy constraints. Additionally, the network typically operates in an open wireless environment, where the inherent broadcasting nature of wireless channels may introduce security vulnerabilities. To address the diverse service demands in large-scale IoT networks and ensure secure information transmission, this study proposes an RIS-enabled secure transmission scheme for a cooperative Simultaneous Wireless Information and Power Transfer Non-Orthogonal Multiple Access (SWIPT-NOMA) system. Methods The RIS is strategically deployed to assist transmission during both the direct and cooperative transmission stages. The goal is to maximize the secrecy rate of the strong NOMA user, subject to the information rate requirements of the weak NOMA user, the energy harvesting needs of the strong NOMA user, and the base station’s minimum transmission power. To solve this multivariable-coupled, non-convex optimization problem, an alternating iterative optimization algorithm is applied. The algorithm optimizes the base station’s active beamforming, the RIS’s passive beam phase shift matrix in the direct transmission stage, the RIS’s active beam phase shift matrix in the cooperative transmission stage, and the PS coefficient of the strong user. These parameters are iteratively adjusted until convergence is achieved. Results and Discussions The convergence of the algorithm is demonstrated in ( Fig. 3 ). As the number of RIS components increases and the number of iterations grows, the secrecy rate of the strong user (U2) gradually improves until it converges. To evaluate the effectiveness of the proposed scheme, it is compared with several benchmark schemes: (1) The random PS coefficient scheme, where RIS is used in both the direct and cooperative transmission stages, and the PS coefficients for strong user U2 are randomly generated. (2) The random RIS phase shift matrix scheme, where RIS enables both transmission stages, with phase shift matrices for both stages randomly generated. (3) The SDR scheme, in which RIS is used in both transmission stages, and the phase shift matrices are optimized using the SDR method. (4) The RIS-enabled direct transmission scheme, where RIS is used only in the direct transmission stage. The impact of the number of base station antennas on the system’s secrecy rate is shown in (Fig. 4 ), and the effect of the number of RIS components on the secrecy rate is explored in (Fig. 5 ). Compared to the other baseline schemes, the proposed scheme achieves a higher secrecy rate for the strong user.Conclusions This paper addresses the challenge of diverse service requirements for users in future large-scale IoT networks and the security of information transmission by designing a secure transmission scheme for an RIS-enabled cooperative SWIPT-NOMA communication system. RIS assists communication in both the direct and cooperative transmission stages. The secrecy rate of the strong user is maximized while considering the information rate requirements of weak NOMA users, the energy harvesting needs of strong NOMA users, and the base station’s minimum transmission power. The proposed optimization problem is a non-convex, multi-variable problem, which is difficult to solve directly. To address this, the problem is divided into several sub-problems, and the active beamforming of the base station, the passive beam phase shift matrix of the RIS in the direct transmission stage, the active beam phase shift matrix of the RIS in the cooperative transmission stage, and the power splitting coefficient of the strong user are iteratively optimized until convergence. Simulation results demonstrate that the secrecy rate of the proposed scheme outperforms that of the scheme where RIS is enabled only in the direct transmission stage. Compared to other baseline schemes, the proposed scheme further enhances the secrecy rate for strong users. -
1. 引言
物联网的蓬勃发展给6G移动通信系统带来了前所未有的挑战[1]。为满足未来网络的大规模连接和来自用户的更高的服务需求,可支持多个用户共享相同资源的非正交多址接入(Non-Orthogonal Multiple Access, NOMA)技术引起了人们的广泛关注[2,3]。为进一步提高NOMA系统的性能,协作NOMA (Cooperative NOMA, CNOMA)技术将NOMA强用户作为中继来辅助NOMA弱用户的通信,可在提供额外的空间分集的同时,进一步保证系统传输的公平性和可靠性[4]。
为解决物联网终端的能量受限问题[5],人们提出了可基于射频信号同时传输信息和能量的无线携能同传技术(Simultaneous Wireless Information and Power Transfer, SWIPT),该技术无需架设电缆即可实现终端供电[6]。同时,物联网处于开放的无线通信环境下,无线信道固有的广播特性给信息安全带来一定挑战,物理层安全技术可利用无线信道的物理特性来实现信息传输的保密性[7]。为最大限度地提高保密速率,人们提出基于可重构智能超表面(Reconfigurable Intelligent Surface, RIS)[8,9]的多天线波束成形技术,利用空间自由度设计发射波束成形,将传输信号引导到合法用户的同时,降低窃听者的接收信号强度,提高保密速率[10]。
作为一种补充设备,RIS无需修改物理层标准,即可灵活地集成到已有网络中,给网络的设计带来了更多的可能性[11,12]。为更好地满足物联网中多种用户的场景需求,RIS被引入SWIPT-NOMA通信系统中以应对未来网络的挑战,如文献[13]考虑未来大型物联网中待服务用户的异构性和服务需求的多样性,针对功率分割(Power Splitting, PS)用户和时间切换(Time Switching, TS)用户共存的异构场景提出一种功率分配方案;文献[14]提出一种将RIS同时用于提高NOMA性能和SWIPT的无线功率传输效率的新网络架构。针对RIS使能的协作SWIPT-NOMA通信系统,文献[15]考虑将RIS引入下行双用户多输入单输出系统,蜂窝中心用户采用PS-SWIPT技术,将信息中继给蜂窝边缘用户,通过联合优化中心用户功率分割比、中继功率以及基站和RIS波束成形,最大化边缘用户的可达速率;文献[16]将多个RIS用于协作传输阶段,辅助蜂窝中心用户向蜂窝边缘用户传输信息;文献[17]提出一种两阶段RIS使能的协作SWIPT-NOMA场景,将RIS同时用于直接传输阶段和协作传输阶段,充分发挥了RIS技术的优势,提高了通信系统的整体性能。RIS使能的无线通信系统中的物理层安全也得到了广泛研究,如文献[18]研究了RIS使能的SWIPT物联网系统中的物理层安全问题,联合优化干扰机协方差矩阵、基站以及RIS波束成形,以最大化能量收集功率;文献[19]考虑RIS使能的PS-SWIPT系统安全通信问题,通过引入人工噪声,最大限度地提高系统最小保密速率;文献[20]考虑同时存在内部窃听者和外部窃听者的RIS使能的NOMA通信系统,在满足用户服务质量需求的情况下,最小化基站的发射功率;文献[21]考虑RIS辅助的强化学习安全无线体域网传输方案,使协调器能够共同优化传感器加密密钥、发射功率和RIS相移以防止主动窃听;文献[22]提出一种RIS辅助的PS-SWIPT通信系统,通过联合优化基站的波束成形矩阵、RIS无源波束成形和PS功率分配系数来最大化用户的可达保密速率。
在已有的关于RIS使能的协作SWIPT-NOMA通信系统的研究中,大多仅考虑将RIS作用于协作NOMA的直接传输阶段或者协作传输阶段。为更好地满足用户服务需求,扩大系统传输的覆盖范围,提高通信系统的整体性能,可考虑通过合理的位置部署,将RIS同时用于直接传输阶段和协作传输阶段。对于存在窃听用户的场景,亦可通过调整RIS的相移矩阵,增强合法用户的通信速率,并抑制窃听用户通信速率,以保证系统信息传输的安全性。
考虑到未来大型物联网中不同用户服务需求的多样性以及信息传输的安全性,本文提出一种RIS使能的协作SWIPT-NOMA通信系统安全传输方案,通过合理部署RIS的位置,将RIS同时作用于直接传输阶段和协作传输阶段,在满足NOMA弱用户(如蜂窝边缘用户)的信息速率需求、NOMA强用户(如蜂窝中心用户)的能量收集需求和基站最小发射功率的条件下,通过联合优化基站的有源波束成形、RIS的相移矩阵、强用户的功率分割系数来最大化中心强用户的保密速率。为解决该非凸问题,采用交替迭代方法将问题拆分成多个子问题,对基站的有源波束成形、直接传输阶段的RIS无源波束相移矩阵、协作传输阶段的RIS有源波束相移矩阵以及强用户的功率分割系数等进行了多次交替迭代优化,直至算法收敛。仿真结果验证了本文算法的收敛性,且与其它基准方案相比,所提方案可以进一步提高强用户的保密速率。
2. 系统建模
2.1 系统模型
考虑如图1所示的RIS使能的下行协作SWIPT-NOMA通信系统,其中,基站端配置M个发射天线,RIS由N个反射元件组成,存在2个单天线用户和1个单天线窃听用户Ue。基站到用户U1,U2的直连信道受到遮挡,因而考虑通过RIS创建反射链路,以保障通信的正常进行。
假设两个单天线用户具有不同的信道条件,用户U2(文中称强用户)比用户U1(文中称弱用户)具有更好的信道条件,即基站-RIS-U2的等效信道优于基站-RIS-U1的等效信道。为更好地满足用户的服务需求,本系统采用协作NOMA策略进行信息传输,具体传输过程如下:(1)在RIS使能的直接传输阶段,RIS反射来自基站的信号,弱用户U1接收信号,强用户U2进行信息传输和能量收集,窃听者在此阶段对强用户U2的信息进行窃听;(2)在RIS使能的协作传输阶段,强用户U2作为中继将直接传输阶段收集到的能量用于向弱用户U1转发信号。
假设基站和RIS已知所有信道的信道状态信息,且所有信道均为准静态信道。所有信道的路径损耗模型均可用PL(d) = C0(d/dd0d0)−α表示,其中,C0 = −30 dB为参考距离d0=1 m处的路径损耗,d为路径距离,α为路径损耗指数[12]。对于小尺度衰落,RIS到所有用户的信道、强用户U2到弱用户U1之间的信道均采用瑞利衰落模型,基站到RIS的信道采用莱斯衰落模型[12]。
2.2 RIS使能的直接传输阶段
在RIS使能的直接传输阶段,基站在RIS的辅助下,通过叠加编码将信息传输给弱用户U1和强用户U2。基站发射的叠加编码信号表示为x=w1s1+w2s2,其中,w1,w2∈CM×1分别表示基站端服务于弱用户U1和强用户U2的有源波束成形向量,s1,s2分别表示弱用户U1和强用户U2的目标接收信号。
弱用户U1的接收信号表示为
yDTU1=hHr,1Θ1G(w1s1+w2s2)+n0 (1) 其中,hHr,1∈C1×N表示从RIS到弱用户U1的信道,Θ1=diag(ejθ11,ejθ12,⋯,ejθ1N)∈CN×N表示直接传输阶段RIS的相移矩阵,θ1n(n=1,2,⋯,N)表示直接传输阶段RIS第n个反射单元的相角,G∈CN×M表示基站到RIS的信道,n0∼CN(0,σ20)为加性高斯白噪声。
强用户U2的接收信号表示为
yDTU2=hHr,2Θ1G(w1s1+w2s2)+n0 (2) 其中,hHr,2∈C1×N表示从RIS到强用户U2的信道。强用户U2采用PS接收机,从功率域将接收信号分割成两个不同的功率流,如图2所示,一部分用于信息解码(information decoding),一部分用于能量收集(energy harvesting)[23],并可根据需要调整功率分割系数ρ,实现信息速率和能量收集之间的平衡[24]。
强用户U2收集的能量表示为
EDTU2,EH=12η(1−ρ)(|hHr,2Θ1Gw1|2+|hHr,2Θ1Gw2|2) (3) 其中,0<η≤1表示能量转换效率。强用户U2用于信息解码的接收信号表示为
yDTU2,ID=√ρ(hHr,2Θ1G(w1s1+w2s2)+n0)+z0 (4) 其中,z0∼CN(0,δ20)表示功率分割后在信息解码过程中产生的噪声。根据NOMA原理,为保证强用户U2处的串行干扰消除(Successive Interference Cancellation, SIC)能够顺利执行,强用户U2解码弱用户U1的可达速率需满足SIC解码速率约束,即
RDTU2,U1=12log2(1+ρ|hHr,2Θ1Gw1|2ρ|hHr,2Θ1Gw2|2+ρσ20+δ20)≥y1,min (5) 其中, {y_{1,\min }} 表示弱用户U1的最小信息速率需求。强用户U2去除弱用户U1的信号后,再解码自身信号,其对应的可达速率表示为
R_{{\text{U2}}}^{{\text{DT}}} = \frac{1}{2}{\log _2}\left( {1 + {{\rho {{\left| {{\boldsymbol{h}}_{{\text{r,2}}}^{\text{H}}{{\boldsymbol{\varTheta}} _1}{\boldsymbol{G}}{{\boldsymbol{w}}_2}} \right|}^2}} \mathord{\left/ {\vphantom {{\rho {{\left| {h_{{\text{r,2}}}^{\text{H}}{{\boldsymbol{\varTheta}} _1}G{w_2}} \right|}^2}} {\left( {\rho \sigma _0^2 + \delta _0^2} \right)}}} \right. } {\left( {\rho \sigma _0^2 + \delta _0^2} \right)}}} \right) (6) 窃听用户Ue的接收信号可表示为
y_{{\text{Ue}}}^{{\text{DT}}} = {\boldsymbol{h}}_{{\text{r,e}}}^{\text{H}}{{\boldsymbol{\varTheta}} _1}{\boldsymbol{G}}\left( {{{\boldsymbol{w}}_1}{s_1} + {{\boldsymbol{w}}_2}{s_2}} \right) + {n_0} (7) 其中, {\boldsymbol{h}}_{{\text{r}},{\text{e}}}^{\text{H}} \in {\mathbb{C}^{1 \times N}} 表示从RIS到窃听用户Ue的信道。在直接传输阶段,窃听用户Ue窃听强用户U2的信息,对应的可达速率表示为
\begin{split} R_{{\text{Ue,U2}}}^{{\text{DT}}} =\;& \frac{1}{2}{\log _2}\left(1 +\frac{ {{{\left| {{\boldsymbol{h}}_{{\text{r,e}}}^{\text{H}}{{\boldsymbol{\varTheta}} _1}{\boldsymbol{G}}{{\boldsymbol{w}}_2}} \right|}^2}}} { {\left( {{{\left| {{\boldsymbol{h}}_{{\text{r,e}}}^{\text{H}}{{\boldsymbol{\varTheta}} _1}{\boldsymbol{G}}{{\boldsymbol{w}}_1}} \right|}^2} + \sigma _0^2} \right)} }\right) \end{split} (8) 因此,强用户U2的保密速率可表示为 R_{{\text{U2}},{\text{sec}}}^{{\text{DT}}} = R_{{\text{U2}}}^{{\text{DT}}} - R_{{\text{Ue}},{\text{U}}2}^{{\text{DT}}} 。
2.3 RIS使能的协作传输阶段
在RIS使能的协作传输阶段,强用户U2将直接传输阶段收集到的能量用于向弱用户U1发送解码信号 {s_1} ,发射功率为 P_{{\text{U2}}}^{{\text{CT}}} = 2E_{{\text{U2,EH}}}^{{\text{DT}}} 。
弱用户U1的接收信号可表示为
y_{{\text{U1}}}^{{\text{CT}}} = \sqrt {P_{{\text{U2}}}^{{\text{CT}}}} \left( {f_{2,1}^{\text{H}} + {\boldsymbol{f}}_{{\text{r}},1}^{\text{H}}{{\boldsymbol{\varTheta}} _2}{\boldsymbol{g}}} \right){s_1} + {n_0} (9) 其中, {\boldsymbol{f}}_{2,1}^{\text{H}} \in {\mathbb{C}^{1 \times 1}} 表示强用户U2到弱用户U1的直连信道, {\boldsymbol{f}}_{{\text{r}},1}^{\text{H}} \in {\mathbb{C}^{1 \times N}} 表示从RIS到弱用户U1的信道, {{\boldsymbol{\varTheta}} _2} = {\text{diag}}\left( {{{\text{e}}^{{\text{j}}\theta _1^2}},{{\text{e}}^{{\text{j}}\theta _2^2}}, \cdots ,{{\text{e}}^{{\text{j}}\theta _N^2}}} \right) \in {\mathbb{C}^{N \times N}} 表示协作传输阶段RIS的相移矩阵, \theta _n^2\left( {n = 1,2, \cdots ,N} \right) 表示协作传输阶段RIS第n个反射单元的相角, {\boldsymbol{g}} \in {\mathbb{C}^{N \times 1}} 表示从强用户U2到RIS的信道。
在系统完成直接传输和协作传输之后,弱用户U1将从RIS和强用户U2处接收到的信号进行最大比合并,进而实现解码,其可达速率可表示为
\begin{split} {R_{{\text{U1}}}} =\;& \frac{1}{2}{\log _2}\Biggr( 1 + \frac{{{{\left| {{\boldsymbol{h}}_{{\text{r,1}}}^{\text{H}}{{\boldsymbol{\varTheta}} _1}{\boldsymbol{G}}{{\boldsymbol{w}}_1}} \right|}^2}}}{{{{\left| {{\boldsymbol{h}}_{{\text{r,1}}}^{\text{H}}{{\boldsymbol{\varTheta}} _1}{\boldsymbol{G}}{{\boldsymbol{w}}_2}} \right|}^2} + \sigma _0^2}} \\ & + \frac{{P_{{\text{U2}}}^{{\text{CT}}}{{\left| {{\boldsymbol{f}}_{2,1}^{\text{H}} + {\boldsymbol{f}}_{{\text{r}},1}^{\text{H}}{{\boldsymbol{\varTheta}} _2}{\boldsymbol{g}}} \right|}^2}}}{{\sigma _0^2}} \Biggr) \end{split} (10) 3. 问题描述
本节通过合理部署RIS的位置,在满足弱用户U1服务质量的条件下,联合优化基站的有源波束成形向量 {{\boldsymbol{w}}_1},\;{{\boldsymbol{w}}_2} 、直接传输阶段的RIS无源波束相移矩阵 {{\boldsymbol{\varTheta}} _1} 、协作传输阶段的RIS有源波束相移矩阵 {{\boldsymbol{\varTheta}} _2} 以及强用户的功率分割系数 \rho 来最大化强用户U2的保密速率。
该优化问题可表示为
\qquad\quad {\text{P}}0:\quad \mathop {\max }\limits_{{{\boldsymbol{\varTheta}} _1},{{\boldsymbol{\varTheta}} _2},{{\boldsymbol{w}}_1},{{\boldsymbol{w}}_2},\rho } R_{{\text{U2}},{\text{sec}}}^{{\text{DT}}} (11) \qquad\quad {\text{s}}.{\text{t}}{\text{.}}\quad {\left\| {{{\boldsymbol{w}}_1}} \right\|^2} + {\left\| {{{\boldsymbol{w}}_2}} \right\|^2} \le {P_{\max }} (12) \qquad\quad {R_{{\text{U1}}}} \ge {y_{1,\min }} (13) \qquad\quad R_{{\text{U2,U1}}}^{{\text{DT}}} \ge {y_{1,\min }} (14) \qquad\quad 0 \le \theta _n^1 \le 2\pi ,n = 1,2, \cdots ,N (15) 0 \le \theta _n^2 \le 2\pi ,n = 1,2, \cdots ,N (16) \qquad\quad\;\; 0 \le \rho \le 1 (17) 其中,式(12)表示基站端有源发射波束成形的功率约束,其中, {P_{\max }} 表示基站端最大发射功率;式(13)表示弱用户U1需满足的服务质量需求;式(14)是强用户U2的SIC解码速率约束;式(15)表示RIS使能的直接传输阶段RIS反射单元的相移约束;式(16)表示RIS使能的协作传输阶段RIS反射单元的相移约束;式(17)表示强用户的功率分割系数约束。P0为多变量相互耦合的非凸优化问题,难以直接进行优化求解。
4. 优化问题求解
本节将上述非凸优化问题拆分成4个子问题,并分别进行交替迭代优化求解,直至算法收敛。
4.1 基站有源波束成形优化
本节在已知直接传输阶段的RIS的无源波束相移矩阵 {{\boldsymbol{\varTheta}} _1} 、协作传输阶段的RIS的有源波束相移矩阵 {{\boldsymbol{\varTheta}} _2} 以及强用户的功率分割系数 \rho 的条件下优化基站的有源波束成形向量 {{\boldsymbol{w}}_1},{{\boldsymbol{w}}_2} 。
令 {{\boldsymbol{W}}_1} = {{\boldsymbol{w}}_1}{\boldsymbol{w}}_1^{\text{H}} , {{\boldsymbol{W}}_2} = {{\boldsymbol{w}}_2}{\boldsymbol{w}}_2^{\text{H}} , {\boldsymbol{h}}_1^{\text{H}} = {\boldsymbol{h}}_{{\text{r,}}1}^{\text{H}}{{\boldsymbol{\varTheta}} _1}{\boldsymbol{G}} , {{\boldsymbol{H}}_1} = {{\boldsymbol{h}}_1}{\boldsymbol{h}}_1^{\text{H}} , {\boldsymbol{h}}_2^{\text{H}} = {\boldsymbol{h}}_{{\text{r,2}}}^{\text{H}}{{\boldsymbol{\varTheta}} _1}{\boldsymbol{G}} , {{\boldsymbol{H}}_2} = {{\boldsymbol{h}}_2}{\boldsymbol{h}}_2^{\text{H}} , {\boldsymbol{h}}_{\text{e}}^{\text{H}} = {\boldsymbol{h}}_{{\text{r,e}}}^{\text{H}}{{\boldsymbol{\varTheta}} _1}{\boldsymbol{G}} , {{\boldsymbol{H}}_{\text{e}}} = {{\boldsymbol{h}}_{\text{e}}}{\boldsymbol{h}}_{\text{e}}^{\text{H}} ,对优化问题P0中式(11)的目标函数展开式进行换底运算,并进一步地,令 f = {\boldsymbol{f}}_{2,1}^{\text{H}} + {\boldsymbol{f}}_{{\text{r}},1}^{\text{H}}{{\boldsymbol{\varTheta}} _2}{\boldsymbol{g}} , {\varGamma _1} = {4^{{y_{1,\max }}}} - 1 ,可将基站有源波束成形优化问题表示为
\begin{split} & {\text{P}}1:\quad\mathop {\max }\limits_{{{\boldsymbol{W}}_1},{{\boldsymbol{W}}_2}} \left\{ {\ln \left( {\rho \sigma _0^2 + \delta _0^2 + \rho {\mathrm{Tr}}\left( {{{\boldsymbol{H}}_2}{{\boldsymbol{W}}_2}} \right)} \right)} \right. \\ & \qquad\qquad - \ln \left( {\rho \sigma _0^2 + \delta _0^2} \right) \\ & \qquad\qquad - \ln \left( {{\mathrm{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_1}} \right) + {\mathrm{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_2}} \right) + \sigma _0^2} \right)\\ & \qquad\qquad \left.+ \ln \left( {{\mathrm{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_1}} \right) + \sigma _0^2} \right) \right\} \\[-1pt] \end{split} (18) \quad {\text{s}}.{\text{t}}{\text{.}}\quad {\text{Tr}}\left( {{{\boldsymbol{W}}_1}} \right) + {\text{Tr}}\left( {{{\boldsymbol{W}}_2}} \right) \le {P_{\max }} (19) \quad \frac{{{\mathrm{Tr}}\left( {{{\boldsymbol{H}}_1}{{\boldsymbol{W}}_1}} \right)}}{{{\mathrm{Tr}}\left( {{{\boldsymbol{H}}_1}{{\boldsymbol{W}}_2}} \right) + \sigma _0^2}} + \frac{{P_{{\text{U2}}}^{{\text{CT}}}{{\left| f \right|}^2}}}{{\sigma _0^2}} \ge {\varGamma _1} (20) \quad \frac{{\rho {\mathrm{Tr}}\left( {{{\boldsymbol{H}}_2}{{\boldsymbol{W}}_1}} \right)}}{{\rho {\mathrm{Tr}}\left( {{{\boldsymbol{H}}_2}{{\boldsymbol{W}}_2}} \right) + \rho \sigma _0^2 + \delta _0^2}} \ge {\varGamma _1} (21) \quad {{\boldsymbol{W}}_1} \ge 0,{{\boldsymbol{W}}_2} \ge 0 (22) \quad {\text{rank}}\left( {{{\boldsymbol{W}}_1}} \right) = 1,{\text{rank}}\left( {{{\boldsymbol{W}}_2}} \right) = 1 (23) 其中,式(18)的第3项导致该式非凸,根据文献[25]中的引理1,可令 {\text{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_1}} \right) + {\text{Tr}}\left( {{H_{\text{e}}}{{\boldsymbol{W}}_2}} \right) + \sigma _0^2 = {1 / t} ,将式(18)的第3项转化为
\begin{split} & - \ln \left( {{\mathrm{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_1}} \right) + {\mathrm{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_2}} \right) + \sigma _0^2} \right) \\ & \quad= \mathop {{\text{max}}}\limits_{t > 0} \left\{ - t\left( {{\text{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_1}} \right) + {\text{Tr}}\left( {{\boldsymbol{{H}}_{\text{e}}}{{\boldsymbol{W}}_2}} \right) + \sigma _0^2} \right)\right. \\ & \qquad\left.+ \ln t + 1 \right\} \\[-1pt] \end{split} (24) 式(20)为非凸约束,可令 \beta = \left( {1 - \rho } \right){\left| f \right|^2} , {\tilde {\boldsymbol{H}}_1} = {{{{\boldsymbol{H}}_1}} \mathord{\left/ {\vphantom {{{H_1}} {\sigma _0^2}}} \right. } {\sigma _0^2}} , {\tilde {\boldsymbol{H}}_2} = {{{{\boldsymbol{H}}_2}} \mathord{\left/ {\vphantom {{{H_2}} {\sigma _0^2}}} \right. } {\sigma _0^2}} ,将其转化为[26]
\begin{split} & {{{\mathrm{Tr}}\left( {{{\boldsymbol{H}}_1}{{\boldsymbol{W}}_1}} \right)} \mathord{\left/ {\vphantom {{Tr\left( {{H_1}{W_1}} \right)} {\left( {Tr\left( {{H_1}{W_2}} \right) + 1} \right)}}} \right. } {\left( {{\mathrm{Tr}}\left( {{{\boldsymbol{H}}_1}{{\boldsymbol{W}}_2}} \right) + 1} \right)}} + \beta \left( {\mathrm{Tr}}\left( {{{\boldsymbol{H}}_2}{{\boldsymbol{W}}_1}} \right) \right.\\ & \quad \left.+ {\mathrm{Tr}}\left( {{{\boldsymbol{H}}_2}{{\boldsymbol{W}}_2}} \right)\right) \ge {\varGamma _1}\\[-1pt] \end{split} (25) 再引入辅助变量 {x_1} [27],将其转化为
{\left( {{a^{(i)}}{x_1}} \right)^2} + {\left( {{{{\text{Tr}}\left( {{{\tilde {\boldsymbol{H}}}_1}{{\boldsymbol{W}}_2}} \right)} \mathord{\left/ {\vphantom {{{\text{Tr}}\left( {{{\tilde H}_1}{W_2}} \right)} {{a^{(i)}}}}} \right. } {{a^{(i)}}}}} \right)^2} \le 2{\text{Tr}}\left( {{{\tilde {\boldsymbol{H}}}_1}{{\boldsymbol{W}}_1}} \right) - 2{x_1} (26) \beta \left( {{\text{Tr}}\left( {{{\tilde {\boldsymbol{H}}}_2}{{\boldsymbol{W}}_1}} \right) + {\text{Tr}}\left( {{{\tilde {\boldsymbol{H}}}_2}{{\boldsymbol{W}}_2}} \right)} \right) \ge {\varGamma _1} - {x_1} (27) 其中, {a^{(i)}} 的迭代更新规则为{a^{(i)}} = \sqrt {{{{\text{Tr}}{{({{\tilde {\boldsymbol{H}}}_1}{{\boldsymbol{W}}_2})}^{(i - 1)}}} / {x_1^{(i - 1)}}}} 。式(23)为非凸秩一约束,可采用半正定松弛(Semi-Definite Relaxation, SDR)方法将其松弛[28]。
综上所述,优化问题P1可转化为
\begin{split} & {\text{P}}2:\quad \mathop {\max }\limits_{{{\boldsymbol{W}}_1},{{\boldsymbol{W}}_2},t,{x_1}} \left\{ \ln \left( {\rho \sigma _0^2 + \delta _0^2 + \rho {\text{Tr}}\left( {{{\boldsymbol{H}}_2}{{\boldsymbol{W}}_2}} \right)} \right)\right. \\ & \qquad\qquad\quad - \ln \left( {\rho \sigma _0^2 + \delta _0^2} \right) - t\left( {\text{Tr}}\left( {{\boldsymbol{{H}}_{\text{e}}}{{\boldsymbol{W}}_1}} \right)\right. \\ & \qquad\qquad\quad \left. + {\text{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_2}} \right) + \sigma _0^2 \right) + \ln t \\ & \qquad\qquad\quad \left. + 1+ \ln \left( {{\text{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_1}} \right) + \sigma _0^2} \right) \right\},\\ &{\mathrm{s.t}}. 式(19)、式(26)、式(27)、式(21)、式(22)\\[-1pt] \end{split} (28) 该问题为凸问题,可使用交替迭代算法求解,算法详细步骤如算法1所示。
表 1 基站有源波束成形子问题求解算法令 i = 0 ,初始化可行解 {\boldsymbol{W}}_1^{(i)},\;{\boldsymbol{W}}_2^{(i)} ,辅助变量 x_1^{(i)} for 交替更新求解 已知 {\boldsymbol{W}}_1^{(i)},\;{\boldsymbol{W}}_2^{(i)} ,计算 {t^{(i + 1)}} 已知 {\boldsymbol{W}}_1^{(i)},\;{\boldsymbol{W}}_2^{(i)},\;x_1^{(i)} , 计算 {a^{(i + 1)}} 已知 {t^{(i + 1)}},{a^{(i + 1)}} ,求解问题P2,获得{\boldsymbol{W}}_1^{(i + 1)},\;{\boldsymbol{W}}_2^{(i + 1)} ,
x_1^{(i + 1)}更新 i = i + 1 end for:收敛 输出基站有源波束成形矩阵 {\boldsymbol{W}}_1^{(i)},\;{\boldsymbol{W}}_2^{(i)} ,辅助变量 x_1^{(i)} 4.2 直接传输阶段RIS无源波束成形优化
本节将在已知基站的有源波束成形向量 {{\boldsymbol{w}}_1}, {{\boldsymbol{w}}_2} 、协作传输阶段的RIS的有源波束相移矩阵 {{\boldsymbol{\varTheta}} _2} 以及强用户的功率分割系数 \rho 的条件下,优化直接传输阶段的RIS的无源波束相移矩阵 {{\boldsymbol{\varTheta}} _1} 。
令 {\boldsymbol{H}}_{{\text{r,1}}}^{\text{H}} = [{\mathrm{diag}}({\boldsymbol{h}}_{{\text{r,1}}}^{\text{H}}){\boldsymbol{G}},0] , {\boldsymbol{H}}_{{\text{r,2}}}^{\text{H}} = [{\mathrm{diag}}({\boldsymbol{h}}_{{\text{r,2}}}^{\text{H}}){\boldsymbol{G}},0] , {\boldsymbol{H}}_{{\text{r,e}}}^{\text{H}} = [{\mathrm{diag}}({\boldsymbol{h}}_{{\text{r,e}}}^{\text{H}}){\boldsymbol{G}},0] ,引入辅助变量 \bar {\boldsymbol{q}}_1^{\text{H}} = [{{\boldsymbol{q}}_1},1] , {{\boldsymbol{Q}}_1} = {\bar {\boldsymbol{q}}_1}\bar {\boldsymbol{q}}_1^{\text{H}} ,则有
{\left| {{\boldsymbol{h}}_{{\text{r}},1}^{\text{H}}{{\boldsymbol{\varTheta}} _1}{\boldsymbol{G}}{{\boldsymbol{w}}_1}} \right|^2}{\text{=}}{\left| {\bar {\boldsymbol{q}}_1^{\text{H}}{{\boldsymbol{H}}_{{\text{r}},1}}{{\boldsymbol{w}}_1}} \right|^2}{\text{=Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},1}}{{\boldsymbol{W}}_1}{\boldsymbol{H}}_{{\text{r}},1}^{\text{H}}} \right) (29) 优化问题P0中类似的表达式可作类似转换,在此不一一赘述。参考优化问题P0到P1的转化过程,将直接传输阶段RIS无源波束成形优化子问题表示为
\begin{split} & {\text{P}}3:\quad\mathop {\max }\limits_{{{\boldsymbol{Q}}_1}} \left\{ \ln \left( {\rho \sigma _0^2 + \delta _0^2 + \rho {\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},2}}{{\boldsymbol{W}}_2}{\boldsymbol{H}}_{{\text{r}},2}^{\text{H}}} \right)} \right)\right. \\ & \qquad\qquad - \ln \left( {\rho \sigma _0^2 + \delta _0^2} \right) - \ln \left( {\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},{\text{e}}}}{{\boldsymbol{W}}_1}{\boldsymbol{H}}_{{\text{r}},{\text{e}}}^{\text{H}}} \right)\right. \\ & \qquad\qquad\left. + {\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},{\text{e}}}}{{\boldsymbol{W}}_2}{\boldsymbol{H}}_{{\text{r}},{\text{e}}}^{\text{H}}} \right) + \sigma _0^2 \right) \\ & \left.\qquad\qquad + \ln \left( {{\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},{\text{e}}}}{{\boldsymbol{W}}_1}{\boldsymbol{H}}_{{\text{r}},{\text{e}}}^{\text{H}}} \right) + \sigma _0^2} \right) \right\} \\[-1pt] \end{split} (30) {\text{s}}{\text{.t}}{\text{.}}\quad\frac{{{\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},1}}{{\boldsymbol{W}}_1}{\boldsymbol{H}}_{{\text{r}},1}^{\text{H}}} \right)}}{{{\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},1}}{{\boldsymbol{W}}_2}{\boldsymbol{H}}_{{\text{r}},1}^{\text{H}}} \right) + \sigma _0^2}} + \frac{{P_{{\text{U2}}}^{{\text{CT}}}{{\left| f \right|}^2}}}{{\sigma _0^2}} \ge {\varGamma _1} (31) \frac{{\rho {\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},2}}{{\boldsymbol{W}}_1}{\boldsymbol{H}}_{{\text{r}},2}^{\text{H}}} \right)}}{{\rho {\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},2}}{{\boldsymbol{W}}_2}{\boldsymbol{H}}_{{\text{r}},2}^{\text{H}}} \right) + \rho \sigma _0^2 + \delta _0^2}} \ge {\varGamma _1} (32) {{\boldsymbol{Q}}_1} \ge 0,{\left[ {{{\boldsymbol{Q}}_1}} \right]_{nn}} = 1,n = 1,2, \cdots ,N + 1 (33) {\text{rank}}\left( {{{\boldsymbol{Q}}_1}} \right) = 1 (34) 其中,可参考式(18)的转化过程将式(30)转化为凸形式;式(31)为非凸约束,可参考式(20)的转化过程引入辅助变量 {x_2} ,将其转化为
\begin{split} {\left( {{b^{(i)}}{x_2}} \right)^2} & + {\left( {{{{\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\tilde {\boldsymbol{H}}}_{{\text{r}},1}}{{\boldsymbol{W}}_2}\tilde {\boldsymbol{H}}_{{\text{r}},1}^{\text{H}}} \right)} \mathord{\left/ {\vphantom {{{\text{Tr}}\left( {{Q_1}{{\tilde H}_{{\text{r}},1}}{W_2}\tilde H_{{\text{r}},1}^{\text{H}}} \right)} {{b^{(i)}}}}} \right. } {{b^{(i)}}}}} \right)^2} \\ & \le 2{\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\tilde {\boldsymbol{H}}}_{{\text{r}},1}}{{\boldsymbol{W}}_1}\tilde {\boldsymbol{H}}_{{\text{r}},1}^{\text{H}}} \right) - 2{x_2} \end{split} (35) \begin{split} & \beta \left( {{\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\tilde {\boldsymbol{H}}}_{{\text{r}},2}}{{\boldsymbol{W}}_1}\tilde {\boldsymbol{H}}_{{\text{r}},2}^{\text{H}}} \right){\text{ + Tr}}\left( {{{\boldsymbol{Q}}_1}{{\tilde {\boldsymbol{H}}}_{{\text{r}},2}}{{\boldsymbol{W}}_2}\tilde {\boldsymbol{H}}_{{\text{r}},2}^{\text{H}}} \right)} \right)\\ & \qquad\qquad \ge {\varGamma _1} - {x_2} \end{split} (36) 其中, {b^{(i)}} 的迭代更新规则与 {a^{(i)}} 一致。式(34)为非凸秩一约束,采用连续秩一约束松弛方法进行求解[27],使用式(37)将其替换
{{\boldsymbol{u}}_{\max ,1}}{\left( {{\boldsymbol{Q}}_1^{(i)}} \right)^{\text{H}}}{{\boldsymbol{Q}}_1}{{\boldsymbol{u}}_{\max ,1}}\left( {{\boldsymbol{Q}}_1^{(i)}} \right) \ge \alpha _1^{(i)}{\text{Tr}}\left( {{{\boldsymbol{Q}}_1}} \right) (37) 其中, {{\boldsymbol{u}}_{\max ,1}}\left( {{\boldsymbol{Q}}_1^{(i)}} \right) 表示 {\boldsymbol{Q}}_1^{(i)} 的最大特征值对应的最大特征向量, {\boldsymbol{Q}}_1^{(i)} 表示在第i次迭代中 \alpha _1^{(i)} 所对应的优化解。通过在每次迭代后将参数 \alpha _1^{(i)} 从0增加到1,可得局部最优秩一解[29]。
综上所述,优化问题P3可进一步优化成
\begin{split} & {\text{P4}}:\;\; \mathop {\max }\limits_{z,{x_2},{{\boldsymbol{Q}}_1}} \left\{ \ln \left( {\rho \sigma _0^2 + \delta _0^2 + \rho {\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},2}}{{\boldsymbol{W}}_2}{\boldsymbol{H}}_{{\text{r}},2}^{\text{H}}} \right)} \right) \right. \\ & \qquad\qquad- \ln \left( {\rho \sigma _0^2 + \delta _0^2} \right) - z\left( {\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},{\text{e}}}}{{\boldsymbol{W}}_1}{\boldsymbol{H}}_{{\text{r}},{\text{e}}}^{\text{H}}} \right) \right.\\ & \qquad\qquad \left.+ {\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{2}},{\text{e}}}}{{\boldsymbol{W}}_1}{\boldsymbol{H}}_{{\text{r}},{\text{e}}}^{\text{H}}} \right) + \sigma _0^2 \right) + \ln z + 1 \\ & \qquad\qquad\left.+ \ln \left( {{\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{2}},{\text{e}}}}{{\boldsymbol{W}}_1}{\boldsymbol{H}}_{{\text{r}},{\text{e}}}^{\text{H}}} \right) + \sigma _0^2} \right) \right\}, \\[-1pt] \end{split} (38) s.t. 式(35)、式(36)、式(32)、式(33)、式(37)
其中, {\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r,e}}}}{{\boldsymbol{W}}_1}{\boldsymbol{H}}_{{\text{r,e}}}^{\text{H}}} \right) + {\text{Tr}}\left( {{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r,e}}}}{{\boldsymbol{W}}_1}{\boldsymbol{H}}_{{\text{r,e}}}^{\text{H}}} \right) + \sigma _0^2 = {1 \mathord{\left/ {\vphantom {1 z}} \right. } z} 。该问题为凸问题,可使用交替迭代算法求解,算法详细步骤如算法2所示。
表 2 直接传输阶段RIS无源波束成形优化子问题求解算法令 i = 0 ,初始化可行解 {\boldsymbol{Q}}_1^{(i)} ,辅助变量 x_2^{(i)} for 交替更新求解 已知 {\boldsymbol{Q}}_1^{(i)} ,计算 {z^{(i + 1)}} 已知 {\boldsymbol{Q}}_1^{(i)},\;x_2^{(i)} , 计算 {b^{(i + 1)}} 已知 {z^{(i + 1)}},{b^{(i + 1)}} ,求解问题P4,获得 {\boldsymbol{Q}}_1^{(i + 1)},x_2^{(i + 1)} 更新 i = i + 1 end for:收敛 输出基站有源波束成形矩阵 {\boldsymbol{Q}}_1^{(i)} ,辅助变量 x_2^{(i)} 4.3 协作传输阶段RIS无源波束成形优化
本节在已知基站的有源波束成形向量 {{\boldsymbol{w}}_1},{{\boldsymbol{w}}_2} 、直接传输阶段的RIS的有源波束相移矩阵 {{\boldsymbol{\varTheta}} _1} 以及强用户的功率分割系数 \rho 的条件下优化协作传输阶段的RIS的无源波束相移矩阵 {{\boldsymbol{\varTheta}} _2} 。
令 {{\boldsymbol{F}}}_{\text{r},1}^{\text{H}}=[{\mathrm{diag}}({{\boldsymbol{f}}}_{\text{r},1}^{\text{H}}){\boldsymbol{g}},{{\boldsymbol{f}}}_{2,1}^{\text{H}}] ,引入辅助变量 \bar {\boldsymbol{q}}_2^{\text{H}} = [{{\boldsymbol{q}}_2},1] , {{\boldsymbol{Q}}_2} = {\bar {\boldsymbol{q}}_2}\bar {\boldsymbol{q}}_2^{\text{H}} ,则有
{\left| {{\boldsymbol{f}}_{2,1}^{\text{H}} + {\boldsymbol{f}}_{{\text{r,1}}}^{\text{H}}{{\boldsymbol{\varTheta}} _2}g} \right|^2}{\text{ = }}{\left| {\bar {\boldsymbol{q}}_2^{\text{H}}{{\boldsymbol{F}}_{{\mathrm{r}},1}}} \right|^2}{\text{ = Tr}}\left( {{{\boldsymbol{Q}}_2}{{\boldsymbol{F}}_{{\text{r}},1}}{\boldsymbol{F}}_{{\text{r}},1}^{\text{H}}} \right) (39) 协作传输阶段RIS无源波束成形优化子问题可表示为
{\text{P}}5:\;{\text{find}}\quad {{\boldsymbol{Q}}_2} (40) \begin{split} \qquad & \text{s}\text{.t}\text{.}\;\;\frac{\text{Tr}\left({{\boldsymbol{Q}}}_{1}{{\boldsymbol{H}}}_{{\mathrm{r}},1}{{\boldsymbol{W}}}_{1}{\boldsymbol{{H}}}_{\text{r,}1}^{\text{H}}\right)}{\text{Tr}\left({{\boldsymbol{Q}}}_{1}{{\boldsymbol{H}}}_{{\mathrm{r}},1}{{\boldsymbol{W}}}_{2}{{\boldsymbol{H}}}_{\text{r}\text{,}1}^{\text{H}}\right)+{s }_{0}^{2}}\\ & \qquad\quad\; +\frac{{P}_{\text{U}2}^{\text{CT}}\text{Tr}\left({{\boldsymbol{Q}}}_{2}{{\boldsymbol{F}}}_{{\mathrm{r}},1}{{\boldsymbol{F}}}_{\text{r,}1}^{\text{H}}\right)}{{\sigma }_{0}^{2}}\ge {\varGamma }_{1} \end{split} (41) {{\boldsymbol{Q}}_2} \ge 0,{\left[ {{{\boldsymbol{Q}}_2}} \right]_{nn}} = 1,n = 1,2, \cdots ,N + 1 (42) {\text{rank}}\left( {{{\boldsymbol{Q}}_2}} \right) = 1 (43) 这里, P_{{\text{U}}2}^{{\text{CT}}} = \eta (1 - \rho )({\text{Tr(}}{{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},2}}{{\boldsymbol{W}}_1}{\boldsymbol{H}}_{{\text{r,2}}}^{\text{H}}) + {\mathrm{Tr}}({{\boldsymbol{Q}}_1}{{\boldsymbol{H}}_{{\text{r}},2}} {{\boldsymbol{W}}_2}{\boldsymbol{H}}_{{\text{r,2}}}^{\text{H}}{\text{)}}) 。式(43)为非凸秩一约束,采用连续秩一约束松弛方法求解[27],使用式(44)将其替换
{{\boldsymbol{u}}_{\max ,2}}{\left( {{\boldsymbol{Q}}_2^{(i)}} \right)^{\text{H}}}{{\boldsymbol{Q}}_2}{{\boldsymbol{u}}_{\max ,2}}\left( {{\boldsymbol{Q}}_2^{(i)}} \right) \ge \alpha _2^{(i)}{\text{Tr}}\left( {{{\boldsymbol{Q}}_2}} \right) (44) 其中, {{\boldsymbol{u}}_{\max ,2}}\left( {{\boldsymbol{Q}}_2^{(i)}} \right) 表示 {\boldsymbol{Q}}_2^{(i)} 的最大特征值对应的最大特征向量, {\boldsymbol{Q}}_2^{(i)} 表示在第i次迭代中 \alpha _2^{(i)} 所对应的优化解。通过在每次迭代后将参数 \alpha _2^{(i)} 从0增加到1,可得局部最优秩一解[29]。
综上所述,优化问题P5可转化为
\begin{split} & {\text{P6:}}\;{\text{find}}\;\;{{\boldsymbol{Q}}_2},\\ & {\mathrm{s.t}}. \;式(41)、式(42)、式(44) \end{split} (45) 该优化问题为标准凸问题,可直接使用CVX工具箱进行求解。
4.4 PS功率分割系数优化
本节在已知基站的有源波束成形向量 {{\boldsymbol{w}}_1},\;{{\boldsymbol{w}}_2} 、直接传输阶段的RIS的有源波束相移矩阵 {{\boldsymbol{\varTheta}} _1} 以及协作传输阶段的RIS的无源波束相移矩阵 {{\boldsymbol{\varTheta}} _2} 的条件下优化强用户的功率分割系数 \rho 。
PS功率分割系数优化子问题可表示为
\begin{split} {\text{P7:}}\; \mathop {\max }\limits_\rho & \left\{ \ln \left( {\rho \sigma _0^2 + \delta _0^2 + \rho {\text{Tr}}\left( {{{\boldsymbol{H}}_2}{{\boldsymbol{W}}_2}} \right)} \right) \right. \\ & - \ln \left( {\rho \sigma _0^2 + \delta _0^2} \right) - \ln \left( {\text{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_1}} \right)\right.\\ & \left.\left.+ {\text{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_2}} \right) + \sigma _0^2 \right) + \ln \left( {{\text{Tr}}\left( {{{\boldsymbol{H}}_{\text{e}}}{{\boldsymbol{W}}_1}} \right) + \sigma _0^2} \right) \right\} \end{split} (46) {\text{s}}{\text{.t}}{\text{.}}\quad \frac{{{\text{Tr}}\left( {{{\boldsymbol{H}}_1}{{\boldsymbol{W}}_1}} \right)}}{{{\text{Tr}}\left( {{{\boldsymbol{H}}_1}{{\boldsymbol{W}}_2}} \right) + \sigma _0^2}} + \frac{{P_{{\text{U}}2}^{{\text{CT}}}{{\left| f \right|}^2}}}{{\sigma _0^2}} \ge {\varGamma _1} (47) \qquad\, \frac{{\rho {\text{Tr}}\left( {{{\boldsymbol{H}}_2}{{\boldsymbol{W}}_1}} \right)}}{{\rho {\text{Tr}}\left( {{{\boldsymbol{H}}_2}{{\boldsymbol{W}}_2}} \right) + \rho \sigma _0^2 + \delta _0^2}} \ge {\varGamma _1} (48) \qquad\; 0 \le \rho \le 1 (49) 这里, P_{{\text{U}}2}^{{\text{CT}}} = \eta (1 - \rho )({\text{Tr(}}{{\boldsymbol{H}}_2}{{\boldsymbol{W}}_1}{\text{)+Tr(}}{{\boldsymbol{H}}_2}{{\boldsymbol{W}}_2}{\text{)}}) 。式(46)中的第2项非凸,可采用逐次凸逼近方法获取其上界,将优化问题P7转化为
\begin{split} & {\text{P8:}}\; \mathop {\max }\limits_\rho \left\{ \ln \left( {{\rho ^{(i)}}\sigma _0^2 + \delta _0^2} \right) + {{\sigma _0^2 \left( {\rho - {\rho ^{(i)}}} \right)} /{\left( {{\rho ^{(i)}}\sigma _0^2 + \delta _0^2} \right)}} \right\},\\ &{\mathrm{s.t}}. 式(47)、式(48)、式(49)\\[-1pt] \end{split} (50) 其中, {\rho ^{(i)}} 表示第i次迭代时 \rho 对应的值。该优化问题为标准凸问题,可直接使用CVX工具箱进行求解。
4.5 整体算法步骤及复杂度分析
本文所提方案的整体算法流程如算法3所示。
表 3 基于交替迭代的整体算法令 i = 0 ,初始化可行解 {\boldsymbol{w}}_1^{(i)},{\boldsymbol{w}}_2^{(i)},{\boldsymbol{Q}}_1^{(i)},{\boldsymbol{Q}}_2^{(i)},{\rho ^{(i)}} for 交替更新求解 {{\boldsymbol{w}}_1},{{\boldsymbol{w}}_2},{{\boldsymbol{Q}}_1},{{\boldsymbol{Q}}_2},\rho 已知 {\boldsymbol{Q}}_1^{(i)},{\boldsymbol{Q}}_2^{(i)},{\rho ^{(i)}} ,求解问题P2,获得 {\boldsymbol{w}}_1^{(i + 1)},{\boldsymbol{w}}_2^{(i + 1)} 已知 {\boldsymbol{w}}_1^{(i + 1)},{\boldsymbol{w}}_2^{(i + 1)},{\boldsymbol{Q}}_2^{(i)},{\rho ^{(i)}} ,求解问题P4,获得 {\boldsymbol{Q}}_1^{(i + 1)} 已知 {\boldsymbol{w}}_1^{(i + 1)},{\boldsymbol{w}}_2^{(i + 1)},{\boldsymbol{Q}}_1^{(i + 1)},{\rho ^{(i)}} ,求解问题P6,获得
{\boldsymbol{Q}}_2^{(i + 1)}已知 {\boldsymbol{w}}_1^{(i + 1)},{\boldsymbol{w}}_2^{(i + 1)},{\boldsymbol{Q}}_1^{(i + 1)},{\boldsymbol{Q}}_2^{(i + 1)} ,求解问题P8,获得
{\rho ^{(i + 1)}}更新 i = i + 1 end for:收敛 本文所提算法复杂度由优化问题P2, P4, P6, P8的复杂度决定[30,31]。在每一次迭代中,采用内点法求解问题P2的计算复杂度为 O({I_0}{M^{3.5}}) , {I_0} 为P2的迭代次数;P4和P6都是通过内点法求解松弛的半正定规划问题,对应的计算复杂度为 O(({I_1} + {I_2}) {(N + 1)^{3.5}}) , {I_1} 和 {I_2} 分别为P4和P6的迭代次数;P8的计算复杂度为 O({I_3}) , {I_3} 表示P8的迭代次数。假设算法达到收敛所需的迭代次数为 {I_4} ,则所提方案的整体计算复杂度可以表示为 O({I_4}({I_0}{M^{3.5}} + ({I_1} + {I_2}) {(N + 1)^{3.5}} + {I_3})) 。
5. 性能分析
本节通过matlab仿真展示所提方案的性能。假设基站和RIS分别位于(0, 0) m, (20, 20) m,其中基站天线数M=16,RIS元件数N=40。假设弱用户U1、强用户U2和窃听用户Ue分别位于(40, 0) m, (20, 0) m, (10, 30) m。基站到RIS的路径损耗指数为1.6,RIS到所有用户的损耗指数为2.2,强用户U2到弱用户U1的路径损耗指数为3.8。假设弱用户U1的信息速率需求为 {\gamma _{1,\max }} = - 10 dB ,能量转换效率为 \eta = 0.75 ,噪声功率 \sigma _0^2 = - 90 dBm,强用户U2信息解码的额外噪声功率 \delta _0^2 = - 80 dBm。为在性能和计算复杂度之间取得平衡,所提算法的收敛阈值为 \varepsilon = {10^{ - 3}} 。性能分析均采用蒙特卡罗仿真,即100次信道实现求均值。
图3对本文算法的收敛性进行了验证。图3中显示,不同RIS元件个数下,随着迭代次数的增加,强用户U2的保密速率逐渐增大,直至收敛;增大RIS元件个数,也能增大强用户U2的保密速率,提高系统信息传输的安全性;在不同RIS元件数目下,算法在第6次迭代时均基本收敛。
为验证本文所提方案的有效性,将所提方案与多种基准方案对比,基准方案包括:(1)随机PS系数方案,将RIS同时用于使能直接传输阶段和协作传输阶段,强用户U2的PS系数随机生成。(2)随机RIS相移矩阵方案,将RIS同时用于使能直接传输阶段和协作传输阶段,两阶段的RIS相移矩阵均随机生成。(3)SDR方案,将RIS同时用于使能直接传输阶段和协作传输阶段,两阶段的RIS相移矩阵均采用SDR[28]方法进行优化。(4)RIS仅使能直接传输阶段的方案,与文献[15]所提方案类似,将RIS仅用于直接传输阶段,RIS在协作传输阶段并不发挥作用。
图4探讨了基站端天线数目对系统保密速率的影响。由图4可见,在不同方案下,增大基站端天线数,系统的保密速率均逐渐增大。这是因为通过优化基站端的有源波束成形,将传输的信号引导到了合法用户处,同时降低了窃听者的接收信号,从而使得保密速率提高。
图5探讨了RIS元件个数对系统保密速率的影响。由图5可见,在不同的方案中,系统保密速率均随着RIS元件数目的增多而逐渐增大。这是因为通过调整RIS的相移矩阵,增强了合法用户的通信速率,并抑制了窃听用户的通信速率,从而保证了系统信息传输的安全性。
从图4和图5中均可见,本文所提方案的性能优于随机PS系数方案,这是因为PS用户可通过调整PS系数合理有效地控制用户信息解码和能量收集的比重,从而更好地提高系统保密速率。本文所提方案和SDR方案的性能均优于随机RIS相移矩阵方案,这是因为通过优化RIS的相移矩阵,也增强了合法用户的通信速率,抑制了窃听用户的通信速率,从而保证了系统信息传输的安全性;此外,本文所提方案的性能略优于SDR方案,由此可见本文采用连续秩一约束松弛方法求解RIS相移矩阵的优越性。本文所提方案的性能优于RIS仅使能直接传输阶段的方案,很好地凸显了协作传输阶段中RIS所发挥的作用,说明通过合理部署RIS的位置,可使RIS同时使能直接传输和协作传输两个阶段,有利于发挥RIS技术的优势,更好地满足系统内用户的服务需求。
6. 结论
考虑未来大型物联网中不同用户服务需求的多样性以及信息传输的安全性,本文借助协作通信的思想,提出一种RIS辅助的协作SWIPT-NOMA通信系统安全传输方案,通过合理部署RIS的位置,将RIS同时用于辅助直接传输阶段和协作传输阶段,在满足NOMA弱用户信息速率需求、NOMA强用户能量收集需求和基站最小发射功率的条件下,最大化强用户的保密速率,并使用交替迭代的方法解决该非凸优化问题。仿真结果显示了所提方案的优越性。
-
1 基站有源波束成形子问题求解算法
令 i = 0 ,初始化可行解 {\boldsymbol{W}}_1^{(i)},\;{\boldsymbol{W}}_2^{(i)} ,辅助变量 x_1^{(i)} for 交替更新求解 已知 {\boldsymbol{W}}_1^{(i)},\;{\boldsymbol{W}}_2^{(i)} ,计算 {t^{(i + 1)}} 已知 {\boldsymbol{W}}_1^{(i)},\;{\boldsymbol{W}}_2^{(i)},\;x_1^{(i)} , 计算 {a^{(i + 1)}} 已知 {t^{(i + 1)}},{a^{(i + 1)}} ,求解问题P2,获得{\boldsymbol{W}}_1^{(i + 1)},\;{\boldsymbol{W}}_2^{(i + 1)} ,
x_1^{(i + 1)}更新 i = i + 1 end for:收敛 输出基站有源波束成形矩阵 {\boldsymbol{W}}_1^{(i)},\;{\boldsymbol{W}}_2^{(i)} ,辅助变量 x_1^{(i)} 2 直接传输阶段RIS无源波束成形优化子问题求解算法
令 i = 0 ,初始化可行解 {\boldsymbol{Q}}_1^{(i)} ,辅助变量 x_2^{(i)} for 交替更新求解 已知 {\boldsymbol{Q}}_1^{(i)} ,计算 {z^{(i + 1)}} 已知 {\boldsymbol{Q}}_1^{(i)},\;x_2^{(i)} , 计算 {b^{(i + 1)}} 已知 {z^{(i + 1)}},{b^{(i + 1)}} ,求解问题P4,获得 {\boldsymbol{Q}}_1^{(i + 1)},x_2^{(i + 1)} 更新 i = i + 1 end for:收敛 输出基站有源波束成形矩阵 {\boldsymbol{Q}}_1^{(i)} ,辅助变量 x_2^{(i)} 3 基于交替迭代的整体算法
令 i = 0 ,初始化可行解 {\boldsymbol{w}}_1^{(i)},{\boldsymbol{w}}_2^{(i)},{\boldsymbol{Q}}_1^{(i)},{\boldsymbol{Q}}_2^{(i)},{\rho ^{(i)}} for 交替更新求解 {{\boldsymbol{w}}_1},{{\boldsymbol{w}}_2},{{\boldsymbol{Q}}_1},{{\boldsymbol{Q}}_2},\rho 已知 {\boldsymbol{Q}}_1^{(i)},{\boldsymbol{Q}}_2^{(i)},{\rho ^{(i)}} ,求解问题P2,获得 {\boldsymbol{w}}_1^{(i + 1)},{\boldsymbol{w}}_2^{(i + 1)} 已知 {\boldsymbol{w}}_1^{(i + 1)},{\boldsymbol{w}}_2^{(i + 1)},{\boldsymbol{Q}}_2^{(i)},{\rho ^{(i)}} ,求解问题P4,获得 {\boldsymbol{Q}}_1^{(i + 1)} 已知 {\boldsymbol{w}}_1^{(i + 1)},{\boldsymbol{w}}_2^{(i + 1)},{\boldsymbol{Q}}_1^{(i + 1)},{\rho ^{(i)}} ,求解问题P6,获得
{\boldsymbol{Q}}_2^{(i + 1)}已知 {\boldsymbol{w}}_1^{(i + 1)},{\boldsymbol{w}}_2^{(i + 1)},{\boldsymbol{Q}}_1^{(i + 1)},{\boldsymbol{Q}}_2^{(i + 1)} ,求解问题P8,获得
{\rho ^{(i + 1)}}更新 i = i + 1 end for:收敛 -
[1] PRASAD R, MANTRI D S, PANDEY S K, et al. 6G Connectivity-Systems, Technologies, and Applications: Digitalization of New Technologies, 6G and Evolution[M]. Aalborg: River Publishers, 2024: 139–156. [2] SARKAR D, YOGITA, YADAV S S, et al. A comprehensive survey on IRS-assisted NOMA-based 6G wireless network: Design perspectives, challenges and future directions[J]. IEEE Transactions on Network and Service Management, 2024, 21(2): 2539–2562. doi: 10.1109/TNSM.2023.3348138. [3] AL-DWEIK A, ALSUSA E, DOBRE O A, et al. Multi-symbol rate NOMA for improving connectivity in 6G communications networks[J]. IEEE Communications Magazine, 2024, 62(7): 118–123. doi: 10.1109/MCOM.001.2300351. [4] DINH P, ARFAOUI M A, SHARAFEDDINE S, et al. A low-complexity framework for joint user pairing and power control for cooperative NOMA in 5G and beyond cellular networks[J]. IEEE Transactions on Communications, 2020, 68(11): 6737–6749. doi: 10.1109/TCOMM.2020.3009262. [5] BENBUK A A, KOUZAYHA N, COSTANTINE J, et al. Charging and wake-up of IoT devices using harvested RF energy with near-zero power consumption[J]. IEEE Internet of Things Magazine, 2023, 6(1): 162–167. doi: 10.1109/IOTM.001.2200202. [6] SHERAZI H H R, ZORBAS D, and O’FLYNN B. A comprehensive survey on RF energy harvesting: Applications and performance determinants[J]. Sensors, 2022, 22(8): 2990. doi: 10.3390/s22082990. [7] JAMEEL F, WYNE S, KADDOUM G, et al. A comprehensive survey on cooperative relaying and jamming strategies for physical layer security[J]. IEEE Communications Surveys & Tutorials, 2019, 21(3): 2734–2771. doi: 10.1109/COMST.2018.2865607. [8] 卢汉成, 王亚正, 赵丹, 等. 智能反射表面辅助的无线通信系统的物理层安全综述[J]. 通信学报, 2022, 43(2): 171–184. doi: 10.11959/j.issn.1000−436x.2022025.LU Hancheng, WANG Yazheng, ZHAO Dan, et al. Survey of physical layer security of intelligent reflecting surface-assisted wireless communication systems[J]. Journal on Communications, 2022, 43(2): 171–184. doi: 10.11959/j.issn.1000−436x.2022025. [9] FU Min, MEI Weidong, and ZHANG Rui. Multi-active/passive-IRS enabled wireless information and power transfer: Active IRS deployment and performance analysis[J]. IEEE Communications Letters, 2023, 27(8): 2217–2221. doi: 10.1109/LCOMM.2023.3287573. [10] CUI Miao, ZHANG Guangchi, and ZHANG Rui. Secure wireless communication via intelligent reflecting surface[J]. IEEE Wireless Communications Letters, 2019, 8(5): 1410–1414. doi: 10.1109/LWC.2019.2919685. [11] FANG Xingchen. Joint channel estimation algorithm for IRS-assisted multi-user MIMO systems[J]. IEEE Communications Letters, 2024, 28(2): 367–371. doi: 10.1109/LCOMM.2023.3346914. [12] DI RENZO M, ZAPPONE A, DEBBAH M, et al. Smart radio environments empowered by reconfigurable intelligent surfaces: How it works, state of research, and the road ahead[J]. IEEE Journal on Selected Areas in Communications, 2020, 38(11): 2450–2525. doi: 10.1109/JSAC.2020.3007211. [13] 季薇, 刘子卿. IRS辅助的异构SWIPT-NOMA系统资源分配方案[J]. 通信学报, 2024, 45(4): 39–53. doi: 10.11959/j.issn.1000-436x.2024027.JI Wei and LIU Ziqing. Resource allocation scheme for IRS assisted heterogeneous SWIPT-NOMA system[J]. Journal on Communications, 2024, 45(4): 39–53. doi: 10.11959/j.issn.1000-436x.2024027. [14] LI Zhendong, CHEN Wen, WU Qingqing, et al. Joint beamforming design and power splitting optimization in IRS-assisted SWIPT NOMA networks[J]. IEEE Transactions on Wireless Communications, 2022, 21(3): 2019–2033. doi: 10.1109/TWC.2021.3108901. [15] GOKTAS M B, DURSUN Y, and DING Zhiguo. IRS and SWIPT-assisted full-duplex NOMA for 6G umMTC[J]. IEEE Transactions on Green Communications and Networking, 2023, 7(4): 1957–1970. doi: 10.1109/TGCN.2023.3289505. [16] LIU Qiuyan, LU Manchun, LI Na, et al. Joint beamforming and power splitting optimization for RIS-assited cooperative SWIPT NOMA systems[C]. 2022 IEEE Wireless Communications and Networking Conference, Austin, USA, 2022: 351–356. doi: 10.1109/WCNC51071.2022.9771850. [17] SAMY M, AI-HRAISHAWI H, KISSELEFF S, et al. Multiple RIS-assisted cooperative NOMA with user selection[C]. 2023 IEEE Global Communications Conference, Kuala Lumpur, Malaysia, 2023: 1405–1410. doi: 10.1109/GLOBECOM54140.2023.10437225. [18] 朱政宇, 徐金雷, 孙钢灿, 等. 基于IRS辅助的SWIPT物联网系统安全波束成形设计[J]. 通信学报, 2021, 42(4): 185–193. doi: 10.11959/j.issn.1000-436x.2021060.ZHU Zhengyu, XU Jinlei, SUN Gangcan, et al. Secure beamforming design for IRS-assisted SWIPT Internet of Things system[J]. Journal on Communications, 2021, 42(4): 185–193. doi: 10.11959/j.issn.1000-436x.2021060. [19] SHEN Yanyan, WANG Chunjie, ZANG Weilin, et al. Outage constrained max-min secrecy rate optimization for IRS-aided SWIPT systems with artificial noise[J]. IEEE Internet of Things Journal, 2024, 11(6): 9814–9828. doi: 10.1109/JIOT.2023.3325362. [20] HAN Hui, CAO Yang, SHENG Min, et al. IRS-aided secure MISO-NOMA networks towards internal and external eavesdropping[C]. 2022 IEEE Global Communications Conference, Rio de Janeiro, Brazil, 2022: 4364–4370. doi: 10.1109/GLOBECOM48099.2022.10000742. [21] XIAO Liang, HONG Siyuan, XU Shiyu, et al. IRS-aided energy-efficient secure WBAN transmission based on deep reinforcement learning[J]. IEEE Transactions on Communications, 2022, 70(6): 4162–4174. doi: 10.1109/TCOMM.2022.3169813. [22] LI Baogang, SI Fuqiang, HAN Dongsheng, et al. IRS-aided SWIPT systems with power splitting and artificial noise[J]. China Communications, 2022, 19(4): 108–120. doi: 10.23919/JCC.2022.04.009. [23] PERERA T D P, JAYAKODY D N K, SHARMA S K, et al. Simultaneous wireless information and power transfer (SWIPT): Recent advances and future challenges[J]. IEEE Communications Surveys & Tutorials, 2018, 20(1): 264–302. doi: 10.1109/COMST.2017.2783901. [24] TANG Jie, LUO Jingci, LIU Mingqian, et al. Energy efficiency optimization for NOMA with SWIPT[J]. IEEE Journal of Selected Topics in Signal Processing, 2019, 13(3): 452–466. doi: 10.1109/JSTSP.2019.2898114. [25] LI Qiang, HONG Mingyi, WAI H T, et al. Transmit solutions for MIMO wiretap channels using alternating optimization[J]. IEEE Journal on Selected Areas in Communications, 2013, 31(9): 1714–1727. doi: 10.1109/JSAC.2013.130906. [26] REN Juanjuan, LEI Xianfu, PENG Zhangjie, et al. RIS-assisted cooperative NOMA with SWIPT[J]. IEEE Wireless Communications Letters, 2023, 12(3): 446–450. doi: 10.1109/LWC.2022.3229843. [27] XU Yanqing, SHEN Chao, DING Zhiguo, et al. Joint beamforming and power-splitting control in downlink cooperative SWIPT NOMA systems[J]. IEEE Transactions on Signal Processing, 2017, 65(18): 4874–4886. doi: 10.1109/TSP.2017.2715008. [28] WU Qingqing and ZHANG Rui. Joint active and passive beamforming optimization for intelligent reflecting surface assisted SWIPT Under QoS Constraints[J]. IEEE Journal on Selected Areas in Communications, 2020, 38(8): 1735–1748. doi: 10.1109/JSAC.2020.3000807. [29] CAO Pan, THOMPSON J, and POOR H V. A sequential constraint relaxation algorithm for rank-one constrained problems[C]. The 25th European Signal Processing Conference, Kos, Greece, 2017: 1060–1064. doi: 10.23919/EUSIPCO.2017.8081370. [30] BEN-TAL A and NEMIROVSKIAEI A S. Lectures on Modern Convex Optimization: Analysis, Algorithms, and Engineering Applications[M]. Philadelphia, USA: Society for Industrial and Applied Mathematics, 2001. [31] BOYD S and VANDENBERGHE L. Convex Optimization[M]. Cambridge: Cambridge University Press, 2004. -