Terahertz (THz) communications are envisioned to be a promising technology for 6G thanks to its broad bandwidth. However, the large path loss, antenna misalignment, and atmospheric influence of THz communications severely deteriorate its reliability. To address this, hybrid automatic repeat request (HARQ) is recognized as an effective technique to ensure reliable THz communications. This paper delves into the performance analysis of HARQ with incremental redundancy (HARQ-IR)-aided THz communications in the presence/absence of blockage. More specifically, the analytical expression of the outage probability of HARQ-IR-aided THz communications is derived, with which the asymptotic outage analysis is enabled to gain meaningful insights, including diversity order, power allocation gain, modulation and coding gain, etc. Then the long term average throughput (LTAT) is expressed in terms of the outage probability based on renewal theory. Moreover, to combat the blockage effects, a multi-hop HARQ-IR-aided THz communication scheme is proposed and its performance is examined. To demonstrate the superiority of the proposed scheme, the other two HARQ-aided schemes, i.e., Type-I HARQ and HARQ with chase combining (HARQ-CC), are used for benchmarking in the simulations. In addition, a deep neural network (DNN) based outage evaluation framework with low computational complexity is devised to reap the benefits of using both asymptotic and simulation results in low and high outage regimes, respectively. This novel outage evaluation framework is finally employed for the optimal rate selection, which outperforms the asymptotic based optimization.

翻译：太赫兹（THz）通信因其宽带宽被视为6G的潜在候选技术，然而，太赫兹通信面临的大路径损耗、天线失准及大气影响严重制约其可靠性。混合自动重传请求（HARQ）被公认为保障太赫兹通信可靠性的有效技术。本文深入研究了在存在/不存在阻塞情况下，采用增量冗余的HARQ（HARQ-IR）辅助的THz通信性能。具体而言，推导了HARQ-IR辅助THz通信中断概率的解析表达式，并基于此实现渐近中断分析以获取关键性能指标，包括分集阶数、功率分配增益、调制与编码增益等。继而基于更新理论，将长期平均吞吐量（LTAT）表示为中断概率函数。为对抗阻塞效应，提出了一种多跳HARQ-IR辅助THz通信方案并评估其性能。为体现所提方案优势，仿真中引入两种对比方案：Type-I HARQ与追尾合并HARQ（HARQ-CC）。此外，提出一种基于深度神经网络（DNN）的低复杂度中断评估框架，该框架分别利用低中断区间的渐近结果与高中断区间的仿真结果。该新型中断评估框架最终应用于最优速率选择，性能优于基于渐近分析的优化方法。