Research On Radar Signal Parameter Estimation And Sorting Method Of The UAV Swarm Collaborative

In modern warfare,unmanned aerial vehicles have the advantages of small size,low cost,flexible control and can avoid casualties.However,with the emergence of more and more cooperative and non-cooperative radiation sources,the electromagnetic environment of modern battlefields is increasingly complex,and traditional UAV radar signal processing algorithms are difficult to solve many problems in radar reconnaissance.At the same time,due to the continuous improvement of the anti-interception ability of target radiation signals and the emergence of new networked radar systems,single radar has been unable to achieve multi-domain detection and obtain target information in all aspects when detecting enemy radars.Therefore,the UAV swarm under distributed coordination to meet the challenges posed by the complexity and variety of modern electronic warfare is required.At the same time,with the rapid development of computer technology and artificial intelligence,in recent years,deep learning has become a research hotspot in the field of radar.Cognitive electronic technology with the ability to quickly perceive and adapt to the electromagnetic environment will become the research direction of future radar technology.In order to solve the limitations of traditional radar signal parameter estimation and signal sorting in modern radar reconnaissance,this thesis studies the radar signal participation and sorting algorithm for the coordination of the UAV swarm,and uses neural network to realize intelligent signal recognition and sorting.In this thesis,the fusion architecture of the UAV swarm under distributed coordination is studied,and the characteristic parameter estimation algorithm of radar signals for the UAV swarm is studied,and the signal sorting algorithm of the UAV swarm is studied.The main work and contributions of this thesis are as follows:1.According to the traditional Multi-reconnaissance receiver information fusion architecture,the radar signal participation and sorting architecture under the coordination of the UAV swarm is given.The matrix equation is constructed by using the received signal model and Doppler information,and the target signal model is recovered by the least squares solution.2.The typical radar signal models are introduced,and the time-frequency characteristics and parameter estimation methods of these signals are studied.For the case where the parameter measurement error of a single UAV is too large,a variety of data fusion algorithms are used to fuse the radar feature parameters extracted by a single UAV to improve the parameter estimation accuracy.Due to the limitations of traditional identification algorithms,deep residual networks are used to train the bispectral features of radar signals to achieve high-precision intra-pulse modulation type identification.3.The preprocessing algorithm of cooperative radar signal sorting for UAV swarm is studied,and the UAV swarm cooperative sorting is realized based on D-S evidence theory.The UAV swarm cooperation can provide a large number of pulse descriptor samples.When there is a priori information,the Bi-LSTM neural network is used to train the pulse descriptor to achieve high-precision radar signal sorting.The effectiveness of all the above algorithms is verified by simulation,and the specific performance of the algorithms is given.