An Individual Identification Method Of Data Radio

Individual radio identification technique is a significant subject in communication confrontation field and has drawn widespread attention from domestic and foreign researchers. Extracting the individual features that attached to the signals and produced from subtle differences among hardware and represent the identity information of the enemy's communication sources, individual identification determine the sources of signals and position information of the sources, and then help the army to track the enemy's significant communication sources.This paper discusses the formation reason of individual features and the expression of actual radio communication signals. This paper analyzes the differences of individual features on difference modulations of the multimodal data radio. Because the individual features change with the modulation types, this paper firstly introduces the identification of modulation types to the individual identification. This paper puts forward the method of extracting different individual features on different modulations and improves the effect of identifying data radio on multimodal state.The radio signal's amplitude changes along with the baseband signal and the frequency domain features is considered stable when the radio signal carries information on its amplitude on time domain with the modulation types of QAM. So the method decomposes the measured signal to the IMFs that contain only one frequency component based on EEMD, and then calculates the singular value of the IMFs matrix, and uses the IMFs to get instantaneous frequency of the signal with Hilbert transform. Then the radios can be classified by the above two transient features on frequency domain. Comparing with the spectrum symmetry algorithm and another instantaneous algorithm, this method provides a better solution on poor aggregation of features, and improves the classification results in low SNR environment.This method extracts individual features on time domain when the radio signal carries information on frequency domain with the modulation types of FSK or MSK. This method first calculates the envelope of the measured signal on time domain with orthogonalcomponent refactoring. And then quantitative the changes of time domain envelope with the kurtosis index and margin index. Comparing with the time domain higher order individual features, this method reduces the complexity and gets an obvious boundary of features classification.Finally, the features of this method are classified by SVM, and the experimental results prove this method and with the help of SVM classifier can achieve good identification results and satisfy the requirements of the project.