Research On Spectrum Situation Generation Technology For Spectrum Map Construction

With the new new technology of technological revolution and industrial transformation driven by emerging technologies such as cloud computing,big data,internet of things,mobile Internet,and artificial intelligence,the volume of new wireless services has exploded.The illegal use of spectrum resources will become a new normal around the world and it will have a profound impact on China’s economic construction and national security.How to make the invisible and intangible radio environment become intuitive and visible is an important part of the research of electromagnetic spectrum.The spectrum map is an abstract expression of the radio environment.It can display the spectrum situation information in an intuitive and multi-dimensional manner,and provide a visual basis for dynamic spectrum access,spectrum sharing and spectrum supervision.The spectrum situation refers to the current state,comprehensive situation and development trend of the electromagnetic spectrum environment.Spectrum situation generation realizes the integration and rendering of spectrum data,which depicts complex spectrum information from time,frequency,emptiness and so on,and presents it in intuitive visualization form.It is the key technology for the construction of spectrum map.In this paper,the spectrum situation generation technology for spectrum map construction is studied systematically,and the topic selection has theoretical significance and application value.The main work of this paper is as follows.1)Design and implement an indoor localization system which can find the abnormal signal.It based on deep learning algorithm to generate the spectrum situation of the indoor.The deep neural network is used to train the data which is in the fingerprint database.The algorithm uses the autoencoder,batch regularization,stochastic gradient descent,and other methods to improve the speed of performance and calculation of the deep neural network.Though the comarision of the simulation,the suitable parameter settings are obtained.Compared with the traditional machine learning algorithm,the algorithm can directly learn and extract the implicit structural features of the data without pre-processing,and can effectively avoid over-fitting,etc.2)Design and build a set of spectrum monitoring platform of Nanjing University of Aeronautics and Astronautics.The spectrum situation generation technology is extended from the indoor to the outdoor,and its application is extended from the local area to the wide area.The system can be divided into three parts: spectrum data acquisition,data integration and processing,visualization.Aiming at the problem of missing data,a kind of matrix reconstruction algorithm based on Augmented Lagrange Multipliers is used to complete the missing data.After that,data that has been completed is integrated into the geographic information system to form the spectrum map of the campus.The spectrum map can visualize the spectrum state of different frequency bands and different time periods of the campus.3)Design and implement a spectrum big data analysis platform based on satellite.The spectrum situation generation technology is extended from outdoor to satellite,from wide area to space.The platform is composed of a satellite operation demonstration module,a spectrum situation generation module,a spectrum situation prediction module,and an interference source positioning module.In this paper,the spectrum situation generation algorithm of satellite is proposed,which maps the complete spectrum data generated by the prediction and prediction to the specific location on the surface of the Earth.Finally,the processed data is integrated into the geographic information system for data visualization analysis and prediction platform construction.Applying big data storage and analysis system to give users a big data level solution.4)Design and build a small and lightweight data acquisition hardware platform that can be carried on a small UAV.It lays a hardware foundation for data acquisitions of spectrum.The platform can be built quickly by using mature modules,which can realize the dynamic monitoring of electromagnetic spectrum in three-dimensional space and can effectively reduce the influence of monitoring signal interruption and multipath.It can improve the effectiveness and reliability of radio monitoring.