电子系博士生李人旺（指导老师：陶梅霞教授），将于2024年6月7日（周五）举行致远荣誉博士生公开报告，具体信息如下：
报告人：李人旺
指导老师：陶梅霞教授
时间：2024年6月7日（周五）10:00 —12:00
地点：上海交通大学闵行校区电信群楼5-303-A大会议室
【报告题目】 智能超表面辅助的通信与感知系统设计和优化
【报告摘要】
智能超表面（Reconfigurable Intelligent Surface, RIS）由于其低成本、易部署和多功能性而成为第六代（Sixth-generation, 6G）无线网络的潜在技术。RIS在无线通信和感知方面都起到了重要的作用。RIS借助其被动波束赋形能力，可以有效提高通信速率、增强感知精度和分辨率，并且实现无间断的通信和感知。因此，RIS在无线通信和感知领域有着广泛的应用。在RIS辅助通信系统中，通过优化设计RIS的反射系数来提高通信速率至关重要。另一方面，在RIS辅助感知系统中，需要设计高精度的感知算法，并借助RIS的测量矩阵来实现高精度的感知。本文从独立设计通信和感知算法开始，最终将两者整合为一个统一的系统，通过RIS的设计与优化来提高系统的通信速率和感知精度。主要贡献和创新可以归纳如下：
第一，针对基于完全信道状态信息的RIS辅助毫米波多输入多输出通信系统，提出了一种基于流形优化的惩罚算法，用于基站和RIS的联合波束赋形设计，从而极大地降低了基站的发射功率。为了降低硬件成本，基站采用了部分连接的混合模拟和数字波束赋形结构。在所有用户的信干噪比约束条件下，通过联合优化基站的混合波束赋形和RIS的响应矩阵，最小化了基站的发射功率。具体而言，通过引入辅助变量和惩罚因子，将复杂的信干噪比约束中的多个优化变量进行解耦，从而转化为简单的子问题进行有效地求解。数值仿真表明，所提出的算法性能优于现有的解决方案，并且RIS在降低基站发射功率中发挥了关键作用。
第二，针对基于部分信道状态信息的RIS辅助毫米波多输入多输出通信系统，分别考虑了窄带和宽带情况，提出了基于黎曼流形的优化算法，用于基站和RIS的联合波束赋形设计，从而极大地提高了系统的遍历可达速率。针对窄带系统，考虑了基站-用户链路被障碍物遮挡的情况，通过受控理论和琴生不等式推导出了遍历可达速率的近似表达式。针对宽带系统，考虑了存在基站-用户链路的情况，通过利用阵列响应矢量的渐近正交性，克服了遍历可达速率的交叉项不是正定的厄米特矩阵的障碍，从而推导出了遍历可达速率的近似表达式。然后，通过联合优化设计基站的发送协方差矩阵和RIS的反射系数最大化了遍历可达速率。数值仿真验证了所提出的优化算法的有效性，其性能可以接近完全信道状态信息下的通信性能。
第三，针对半被动RIS辅助感知系统，提出了一种基于原子范数最小化的估计算法，用于多目标到达角（direction-of-arrival, DoA）的估计，从而实现了高的感知精度和分辨率。半被动RIS由被动反射单元和主动感知单元组成，其中反射单元将信号从基站反射到目标，感知单元根据目标的回波信号估计出到达角。与仅仅依赖RIS感知单元进行到达角估计的现有文献不同，通过原子范数最小化方法充分挖掘了嵌入在RIS反射单元和感知单元中的角度信息。随后，推导出了到达角估计的克拉美罗界（Cram'er-Rao bound, CRB）以评估估计的性能。数值仿真证实了所提出的估计算法相对于参考基准具有卓越的准确性和分辨率性能。
最后，针对RIS辅助通感一体化系统，提出了一种两阶段的通感一体化协议，从而有效地同时实现了通信和感知功能。该协议包括波束扫描和数据传输两个阶段。在第一阶段，利用波束全空间扫描的的特性，同时获得了通信用户的最佳波束并初步估计出了目标的角度。在第二阶段，通过RIS的波束分裂设计形成了两个波束，一个服务通信用户以保证通信速率，一个服务感知目标以进一步提高感知精度。然后，推导出了通信用户的可达速率以及目标角度估计的克拉美罗界和近似均方误差。数值仿真验证了分析结果以及所提出方案的有效性。
【Abstract】
Reconfigurable intelligent surface (RIS) technology has emerged as a prospective candidate for sixth-generation (6G) wireless networks, owing to its cost-effectiveness, ease of deployment, and versatile capabilities. RIS plays a crucial role not only in communication but also in sensing and localization. Leveraging its passive beamforming attributes, RIS facilitates improved communication rate, enhanced sensing accuracy and resolution, and uninterrupted communication and sensing. Consequently, RIS finds widespread applications in wireless communication and sensing domains. In RIS-aided communication systems, the design and optimization of RIS reflection coefficients are crucial for improving communication rates. On the other hand, in RIS-aided sensing systems, the design of sensing algorithms utilizing the RIS measurement matrix is imperative to achieve high sensing accuracy. This thesis begins with the independent design of communication and sensing algorithms, and eventually integrates them into a unified system, with the aim of boosting communication rates and sensing accuracy through RIS design and optimization. The main contributions and innovations are outlined as follows:
Firstly, for RIS-assisted millimeter wave (mmWave) multiple-input multiple-output (MIMO) communication systems based on full channel state information (CSI), a penalty-based algorithm is proposed for joint base station (BS) and RIS beamforming design, significantly reducing the transmit power at the BS. In order to reduce the hardware costs, a sub-connected hybrid analog and digital beamforming structure is employed at the BS. Under the signal-to-interference-plus-noise ratio (SINR) constraints of all users, the transmit power is minimized by jointly optimizing hybrid beamforming at the BS and the response matrix at the RIS. Specifically, by introducing auxiliary variables and a penalty factor, we decouple the optimization variables within the complex SINR constraints, and then transform them into simple sub-problems for efficient solution. Numerical results demonstrate that the proposed algorithm outperforms existing solutions, highlighting the pivotal role of RIS in power reduction.
Secondly, for RIS-aided mmWave MIMO systems based on partial CSI, both narrowband and wideband scenarios are addressed. For narrowband systems, considering the scenario where the BS-user link is obstructed by obstacles, approximate expressions for the ergodic achievable rate are derived by means of the majorization theory and Jensen's inequality. For wideband systems, considering the existence of the BS-user link, the asymptotic orthogonality of the array response vectors is utilized to overcome the hurdle of the non-positive semidefinite Hermitian matrix of the cross product term of the ergodic achievable rate. Subsequently, alternating optimization-based algorithms are proposed to maximize the ergodic achievable rate by jointly designing the transmit covariance matrix at the BS and the reflection coefficients at the RIS. Numerical results validate the effectiveness of the proposed optimization algorithm, which can approach the performance achievable under full CSI.
Thirdly, for semi-passive RIS-aided sensing systems, an atomic norm minimization (ANM)-based estimation algorithm is proposed for multiple target direction-of-arrival (DoA) estimation. The semi-passive RIS consists of passive reflecting elements (REs) and active sensing elements (SEs), where REs reflect signals from the BS to targets, and SEs estimate DoA based on echo signals reflected by the targets. Instead of solely relying on RIS SEs for DoA estimation as done in the existing literature, we fully exploit the DoA information embedded in both RIS REs and SEs via the ANM scheme. Subsequently, the Cram'er-Rao bound (CRB) for DoA estimation is derived to evaluate the estimation performance. Extensive numerical results substantiate the superior accuracy and resolution performance of the proposed algorithm over representative baselines.
Finally, for RIS-aided integrated communication and sensing (ISAC) systems, a novel ISAC protocol is proposed to effectively achieve communication and sensing simultaneously. The protocol consists of two stages: beam scanning and data transmission. In the first stage, the characteristic of beam scanning across the entire space is utilized to obtain the best beam of communication user and initially estimate the target angle simultaneously. In the second stage, two beams are formed by designing RIS beam splitting, where one serves communication users to ensure communication rates and the other serves target sensing to further enhance sensing accuracy. We then derive the achievable rate of the communication user, as well as the CRB and the approximate mean square error of the target angle estimation. Extensive numerical results are provided to verify our analysis as well as the effectiveness of the proposed scheme.
