Research On Distributed Intelligent Cooperative Monitoring Technology Of Space-based Electromagnetic Spectrum

By deploying radio signal transceivers on satellites,the space-based platform can effectively process and monitor electromagnetic spectrum signals emitted by radiation sources by taking full advantage of the unique advantages of being independent of terrain and monitoring visual range.With the wide use of rf terminals,the signal data to be processed by satellite-borne radio equipment increases exponentially.Due to the shortage of satellite storage,computing and network resources,the traditional computing mode can no longer meet the real-time,bandwidth,security and other requirements of spectrum monitoring.How to reduce data transmission,improve real-time performance,reduce resource occupancy,control star clusters and implement distributed collaborative monitoring of spectrum have become the focus of research.Therefore,this thesis applies edge computing and artificial intelligence technology to space-based spectrum monitoring,mainly to identify the modulation mode of the signal,and constructs an intelligent network collaborative recognition system of space-based spectrum signal modulation mode.The main work and innovative achievements are summarized as follows:(1)To solve the problems of large data transmission,heterogeneous satellite cluster and difficult spectrum application deployment,a space-based distributed intelligent collaborative monitoring architecture of electromagnetic spectrum is proposed.Edge computing and artificial intelligence technology are applied to space-based spectrum monitoring.Satellites are regarded as edge nodes.Intelligent spectrum monitoring algorithm is deployed to satellites to form edge intelligent satellite nodes.The edge computing processing framework is used to control heterogeneous distributed satellites in a unified manner,and the cloud delivers spectrum monitoring to edge satellite nodes on demand,so as to realize in-orbit signal processing and inter-satellite collaborative signal recognition.(2)A spectrum monitoring method based on distributed deep learning is proposed to solve the problems of shortage of on-board resources,large number of intelligent monitoring algorithm parameters and huge computing costs caused by massive spectrum data.A lightweight neural network is constructed,and the reasoning model is segmented and deployed on edge satellite nodes.Using intersatellite algorithm and data coordination mechanism,spectrum monitoring task is accomplished jointly.(3)To solve the problems of intelligent monitoring algorithm occupying too much storage resources,complex operating environment and difficult deployment,the container space-based spectrum monitoring method and application deployment strategy are proposed.The algorithm and operating environment are encapsulated in a container,which is conducive to the unified distribution of applications in the constructed spectrum monitoring architecture,and occupies less memory resources,and the startup speed is fast.(4)Based on the constructed monitoring architecture,a set of verification system with distributed intelligent coordination monitoring function of electromagnetic spectrum is built.The cloud adopts servers rich in storage and computing resources and heterogeneous intelligent edge hardware to simulate edge satellite nodes.In addition,intelligent recognition algorithm of signal modulation mode is deployed in container form to realize highly reliable service of spectrum signal modulation mode recognition.The research content of this thesis is based on the existing space-based spectrum monitoring technology and fully absorbs the emerging technology.After theoretical analysis,spectrum monitoring framework is established,intelligent monitoring algorithm is designed and containerized.According to the deployment strategy,spectrum monitoring applications are deployed to edge satellite nodes on demand to realize the coordinated identification of modulation modes.