Power amplifier individual identification method and device

Abstract

The invention provides a power amplifier individual identification method. Different fingerprint features are extracted by utilizing difference of extracted features of an EfficentNets series heterogeneous convolutional neural network under a low signal-to-noise ratio condition, and the fingerprint feature extraction capability is improved by integrating prediction results of a plurality of networks by using a stacking method. The invention further provides a power amplifier individual identification device which comprises a signal receiving module, a digitization module, a data processing module and a power supply and is simple and practical in structure. According to the method, the parameter amount and the floating point calculation amount are small, and the method is suitable for beingtransplanted to embedded hardware equipment such as an FPGA for use; according to the method, the power amplifier of a certain specific radiation source can be accurately identified by receiving thesteady-state signals of the same batch of radiation sources of a non-cooperative party, so that the specific orientation of the specific radiation source is determined through methods such as radio positioning and the like, and the motion trail of the specific radiation source is tracked.

Description

technical field [0001] The invention relates to the technical field of radiation source identification, in particular to an individual identification method and identification device for a steady-state signal power amplifier. Background technique [0002] The non-stationary components of some signals will show cyclostationarity in high-order statistics, which is specifically manifested in that the signal will show single-period or multi-period smooth changes in mean values, correlation functions, and high-order cumulants. These phenomena are more common in artificial signals such as communication and radar, or signals that change periodically according to natural laws. In the nonlinear behavior model of memory effect power amplifiers, the periodicity of high-order statistics is an important feature that can extract fingerprint information from signals. [0003] There are slight differences in the output signals of power amplifiers of the same type due to nonlinear character...

AI Technical Summary

Problems solved by technology

[0005] Since the internal components of the radiation source will cause joint interference to the signal, it is difficult to analyze the formation mechanism of the spurious noise of the radiation source, and it is impossible to establish a mathematical model of fingerprint characteristics for the radiation source as a whole

[0006] The fingerprint features generated during the working process of the radiation source mainly come from the spurious noise generated by the power amplifier. The difference between the fingerprint features of the radiation source is subtle and the recognition rate is easily reduced due to noise interference.

Images