Detection Method Of Small Unmanned Aerial Vehicle(UAV)Based On External Emitter Radar

As the functions of small unmanned aerial vehicle(UAV)become more and more powerful,their applications in various fields are gradually increasing,but they also bring many potential safety hazards.The phenomenon of "black flying" and "indiscriminate flying" endlessly emerges.Therefore,the issue of UAV supervision has become the top priority,and UAV target detection has become the primary task.Radar detection technology is still the most widely used method at present.Based on the detection method of active radar has high equipment cost and serious electromagnetic pollution,and many deficiencies for detecting "low,slow and small" targets such as UAV.For this reason,the UAV detection method based on external radiation source radar proposed in this paper,and focuses on the cyclic spectrum and deep learning network to detect the performance of the UAV micro-motion characteristics in the echo signal.In order to effectively detect UAV targets,this paper conducts in-depth research on the above aspects:(1)The signal propagation model of the external radiation source radar is deduced,and the DTMB standard and signal model are introduced.At the same time,the micro-motion model of the rotor UAV is analyzed,which provides a theoretical basis for the follow-up research.Then the traditional time-frequency conversion method is used to study the time-frequency characteristics of the direct wave signal and UAV target echo signal,and the influence of the direct wave signal on the time-frequency diagram of UAV target echo signal is simulated and analyzed,as well as the anti-noise performance of the time-frequency conversion method.(2)On the basis of the echo signal model of the external radiation source radar,a method of UAV detection based on the cyclic spectrum is proposed.First,the cyclostationary characteristics of each component signal in the model are deduced,and the difference between the UAV echo signal and the rest of the signals in the cyclic spectrum is analyzed,and then the cyclic spectrum contour map of each signal is extracted as the detection target,and the difference of energy distribution on the contour map caused by the above difference was used for UAV detection.The simulation results show that the existence of UAV can be distinguished by comparing the gray value of the contour map,and the purpose of detecting UAV can be achieved without direct wave and noise suppression.The experimental results verify the effectiveness of the method.(3)A detection method for rotor UAV based on Weight Agnostic Neural Networks(WANNs)is proposed.As the expansion of cyclic spectrum detection method of UAV,this method mainly studies the detection performance of UAV in different flight states using neural network.Firstly,the micro-motion model of the rotor UAV is derived,and then the construction process of WANNs model is explained in detail.Then,the cyclic spectral contour map of echo signal is used as the training and testing data set.Finally,the simulation results show that the proposed method is robust to noise.At the same time,the actual measurement results also verify that the WANNs model can effectively identify the micro-motion characteristics of the UAV in various flight states,and has good image recognition capabilities.