Design of Spectrum Monitor Based on Edge Computing

This paper provides a test system for spectrum monitor based on edge computing. The spectrum monitor is mainly composed of software-defined radio (SDR) core board, edge computing terminal, and spectrum monitor processor. The working frequency of spectrum monitor is 100kHz-6GHz, meeting the needs of Sub-6GHz frequency range of 5G communications. The SDR core board preprocesses the received wireless modulation signal and transmits it to the edge computing terminal. The edge computing terminal recognizes the received signal by deploying deep learning network model. Different modulation signals generated by the signal source are used to train and test network model, which can improve the recognition accuracy. The spectrum monitor based on edge computing has the characteristics of small size, low power consumption and low latency. With the test of the different modulation signal, the recognition accuracy of spectrum monitor can reach 94%, which meets the design requirement.