In this letter, we study efficient channel estimation and passive beamforming designs for a double-intelligent reflecting surface (IRS) aided single-user communication system, where a user communicates with an access point (AP) via the cascaded user-IRS 1-IRS 2-AP double-reflection link. First, a general channel estimation scheme is proposed for the system under any arbitrary inter-IRS channel, where all coefficients of the cascaded channel are estimated. Next, for the typical scenario with a line-of-sight (LoS)-dominant inter-IRS channel, we propose another customized scheme to estimate two signature vectors of the rank-one cascaded channel with significantly less channel training time than the first scheme. For the two proposed channel estimation schemes, we further optimize their corresponding cooperative passive beamforming for data transmission to maximize the achievable rate with the training overhead and channel estimation error taken into account. Numerical results show that deploying two cooperative IRSs with the proposed channel estimation and passive beamforming designs achieves significant rate enhancement as compared to the conventional case of single IRS deployment.

中文翻译：

---

通过双 IRS 进行无线通信：信道估计和无源波束成形设计

在这封信中，我们研究了双智能反射面 (IRS) 辅助单用户通信系统的有效信道估计和无源波束成形设计，其中用户通过级联用户 IRS 1-IRS 与接入点 (AP) 通信2-AP双反射链路。首先，针对任意IRS间信道下的系统提出了通用信道估计方案，其中级联信道的所有系数都被估计。接下来，对于具有视距 (LoS) 主导的 IRS 间信道的典型场景，我们提出了另一种定制方案来估计秩为一的级联信道的两个特征向量，其信道训练时间明显少于第一种方案. 对于两个提出的信道估计方案，我们进一步优化了它们对应的用于数据传输的协作无源波束成形，以在考虑到训练开销和信道估计误差的情况下最大化可实现的速率。数值结果表明，与单 IRS 部署的传统情况相比，部署两个协作 IRS 与所提出的信道估计和无源波束成形设计相比，实现了显着的速率增强。