Signal Modulation Mode Recognition Method Research Based On Duffing Oscillator

With the development of communication technology, wireless communication environment becomes increasingly complicated. Thus various modulation mode of signals exist in the limited bandwidth resources. The purposes of modulation recognition is to determine the modulation of these signals, preparing for signal recovery. Modulation recognition plays an important role in military field and civilian field.This paper first introduced the standard form of 2ASK, 2FSK, 2PSK, QPSK, 16 QAM, 32 QAM and 64 QAM. Then researched and analyzed several classic modulation recognition algorithms, and decided to extract under test signal amplitude and phase characteristics as the modulation recognition characteristic parameters. Then we chose Duffing oscillator as the signal preprocessing algorithm based on initial value sensitivity and noise immunity of the chaos theory. Next, we utilized the wave detection algorithm to extract the amplitude of the time domain waveform. We designed a smoothing algorithm because the noise made the oscillator appear slight fluctuations, for making the extracted wave characteristic value smoothing. We put forward the clustering algorithm based on density by studying the principle of clustering algorithm. This algorithm was very suitable for this paper, clustering the amplitude characteristic parameters. We could divide the under test signals into 2ASK and 2FSK, 2PSK and QPSK, 16 QAM, 32 QAM, 64 QAM because each signal’s amplitude characteristic parameters are difference. The next, we designed the modulation mode decision. The distance of the clustering points and zero was treated as 2ASK and 2FSK judgment conditions, and treated the Poincare section number of the fixed point as 2PSK and QPSK judgment conditions. Then we simulated the antinoise performance of the modulation recognition method, and simulated the identification probability affected by frequency offset. These simulations proved this topic modulation recognition method with high recognition probability in low SNR and the effectiveness of the modulation recognition method. Finally, we designd the test interfaces of the modulation recognition method. The experiment of identifying each signal modulation proved that this topic modulation recognition method could correctly identify seven kinds of modulation signal.