Research on Spectrum Intelligent Recognition Technology Based on an Edge Processing Framework KubeEdge

摘要

In recent years, artificial intelligence technology has developed rapidly, spectrum recognition presents the characteristics of intelligence. However, with the increase of the amount of radio spectrum data, problems such as the large amount of computation of artificial intelligence algorithm and the shortage of storage and computing resources of spectrum recognition terminal become more and more prominent, resulting in low efficiency of spectrum recognition. Combined with edge computing technology, an spectrum intelligent recognition framework based on edge processing framework KubeEdge is proposed in this paper. The architecture uses KubeEdge framework to co-manage cloud server and edge spectrum recognition node. The training of spectrum intelligent recognition algorithm is realized in the cloud, and the spectrum of input signals is recognized and classified at the edge node, so as to achieve the purpose of cloud and edge collaborative recognition and improve the efficiency of spectrum recognition.