Use of smart directional antennas in handheld devices to generate a narrow beam in different directions for mmWave/TeraHertz communications present significant challenges. Devices using such antennas may have to scan several different directions in three-dimensional (3D) space to discover another user or an access point, a process that can result in problematic delays. Moreover, small movements of a user/device in the form of rotation and/or displacement may cause the discovered link to be lost. This paper proposes adaptive link discovery algorithms for devices in both infrastructure/ad hoc networks and evaluates their performance in terms of time-to-discovery. We show that one of the two proposed methods provides guaranteed discovery. We use an inertial measurement unit sensor to help intelligently rediscover a lost/degraded link. We propose sensor assisted link prediction methods for low-latency rediscovery in 3D space. We evaluate the effectiveness of our prediction-based rediscovery methods by testing them with real datasets representing various user/device 3D rotation patterns. We show that the smoothing based rediscovery can reach the prediction accuracy to 100% when two antenna sectors are searched, and it reduces the time-to-rediscovery by up to ${S}\text{x}$ ( ${S}$ times) as compared to the time-to-discovery, where ${S}$ is the number of antenna sectors.

中文翻译：

---

具有定向天线的毫米波和太赫兹设备的智能自主用户发现和链路维护

在手持设备中使用智能定向天线在不同方向上为毫米波/太赫兹通信生成窄波束是一项重大挑战。使用此类天线的设备可能必须在三维 (3D) 空间中扫描多个不同方向以发现另一个用户或接入点，这一过程可能会导致延迟问题。此外，用户/设备以旋转和/或位移的形式进行的微小移动可能会导致发现的链接丢失。本文为基础设施/自组织网络中的设备提出了自适应链路发现算法，并评估了它们在发现时间方面的性能。我们表明，所提出的两种方法之一提供了有保证的发现。我们使用惯性测量单元传感器来帮助智能地重新发现丢失/退化的链路。我们提出了传感器辅助链接预测方法，用于 3D 空间中的低延迟重新发现。我们通过使用代表各种用户/设备 3D 旋转模式的真实数据集对其进行测试来评估基于预测的重新发现方法的有效性。我们表明，当搜索两个天线扇区时，基于平滑的重新发现可以达到 100% 的预测精度，并将重新发现时间减少多达 ${S}\text{x}$ ( ${S}$ 次）与发现时间相比，其中 ${S}$ 是天线扇区的数量。