This study deals with the problem of fiber-free optical communication systems—known as free space optics—using received signal strength identifier (RSSI) prediction analysis for hard switching of optical fiber-free link to base radio-frequency (RF) link and back. Adverse influences affecting the atmospheric transmission channel significantly impair optical communications, therefore attention was paid to the practical design, as well as to the implementation of the monitoring device that is used to record and process weather information along a transmission path. The article contains an analysis and methodology of the solution of the high availability of the optical link. Attention was paid to the technique of hard free space optics (FSO)/RF-switching with regard to the amount of received optical power detected and its relation to the quantities influencing the optical communication line. For this purpose, selected methods of machine learning were used, which serve to predict the received optical power. The process of analysis of prediction of received optical power is realized by regression models. The study presents the design of the optimal data input matrix model, which forms the basis for the training of the prediction models for estimating the received optical power.

中文翻译：

---

混合FSO / RF系统的接收光功率预测

这项研究解决了无光纤光通信系统（称为自由空间光学系统）的问题，它使用接收信号强度标识符（RSSI）预测分析来将无光纤链路硬切换为基本射频（RF）链路并返回。影响大气传输通道的不利影响大大损害了光通信，因此，人们对实际设计以及用于沿传输路径记录和处理天气信息的监视设备的实施给予了关注。本文包含光链路高可用性解决方案的分析和方法论。在检测到的接收光功率的数量及其与影响光通信线路的数量之间的关系方面，人们对硬自由空间光学（FSO）/ RF切换技术给予了关注。为此，使用了选定的机器学习方法，这些方法可用来预测接收到的光功率。通过回归模型来实现对接收光功率的预测的分析过程。该研究提出了最佳数据输入矩阵模型的设计，这为训练用于估计接收光功率的预测模型奠定了基础。通过回归模型实现对接收光功率预测的分析过程。该研究提出了最佳数据输入矩阵模型的设计，这为训练用于估计接收光功率的预测模型奠定了基础。通过回归模型来实现对接收光功率的预测的分析过程。该研究提出了最佳数据输入矩阵模型的设计，这为训练用于估计接收光功率的预测模型奠定了基础。