Research On Aircraft Motion Pattern Recognition Based On Support Vector Machine

In the modern electronic warfare, making use of all detection method to acquire useful information becomes more and more important. If we get more adequate information, it will be more helpful to establish strategy. And it will be easier to win the final victory. Meanwhile electronic jamming technology has been widely used in modern warfare, so how to fight against electronic jamming and identify the electronic interference sources, which can guide us to make the right strategy, are of great significance. In order to realize the anti-electronic jamming, we need make further research on related technology of radar signal processing and explore new ways to solve the problem. Because passive radar has the advantages of good concealment, strong survivability and so on, we researched the direct electromagnetic signal from the airborne radar during the flight, which received by a single ground-based passive radar. And then we proposed a method combined time-frequency analysis and support vector machine to process this kind of signal in order to correctly identify the aircraft flight mode. The research content of this paper is only a part of the identification of the electronic countermeasure airplane, which is active jamming. As the forefront theory study, it is mainly aimed at the identification of the aircraft flight pattern of the long distance and the airborne radar with omnidirectional antenna, taking into account that there are the uniform rectilinear flight, uniform circular flight, uniform elliptic runway flight and uniform eight-shaped trajectory flight. The main works are as follows:(1) According to the radar equation and actual aircraft flight speed, electromagnetic carrier wave, three-dimensional space relative position, noise and some other parameters, simulation signals were generated respectively with uniform rectilinear flight, uniform circular flight, uniform elliptic runway flight and uniform eight-shaped trajectory flight.(2) During the flight, due to the change of the radial velocity of the aircraft relative to the passive radar receiving station, the Doppler shift is also changed with time. So we deal with the Doppler shift of different flight modes signal by the method of time-frequency analysis: Hilbert Huang transform(HHT), which has good effect on non- stationary signal.(3) Because under the different flight modes Doppler frequency spectrum has their respective characteristics and ensemble empirical mode decomposition(EEMD) acts essentially as a dyadic filter bank, so the ratio of every Intrinsic Mode Function’s(IMF) energy to total IMF’s energy is extracted as the feature vector. And then we collected a group of sample data to use the support vector machine to train and recognize.This paper researched direct electromagnetic wave signals emitted by airborne radar, and finally obtained a good result when use the method combined with frequency analysis and support vector machine. The correct recognition rate about these four flight pattern is 86.25%. That’s to say the method proposed in this paper provides a new way to Electronic jamming recognition.