Research On Radio Radiation Characteristics Of Space Targets

SNR estimation and modulation classification play an important role in wireless communication,SNR estimation plays an important role in adaptive demodulation,gain control and channel allocation,modulation classification plays an important role in electronic warfare,electromagnetic spectrum monitoring and cognitive radio.Due to the shortage of space spectrum resources and the increasing complexity of communication environment,higher requirements are required for SNR estimation and modulation style classification algorithm.In this paper,non-data-assisted SNR estimation algorithms are simulated and summarized,and the modulation recognition technology based on feature classification is emphatically studied.The SNR estimation algorithms summarized in this paper mainly include: highorder moment estimation algorithm,mean-variance estimation algorithm and signal subspace decomposition estimation algorithm.After the simulation analysis,the estimation algorithm of high order moment and mean variance can only deal with the modulation mode of constant envelope,and the mean variance algorithm has a small mean square error in the SNR estimation under low SNR.SNR estimation algorithm of signal subspace decomposition with mean zero white gaussian noise,the correlation function matrix constructed by the way to estimate signal-to-noise ratio,but for the constant envelope modulation method to estimate signal-to-noise ratio,signal subspace and noise subspace estimation affects the estimation precision of the cut-off point,in this paper,three kinds of estimation computation and accuracy of the algorithm is analyzed and summarized.Modulation style classification is the key research object of this paper.This paper first introduces the transmission model of communication signals,and introduces the common space communication frequency band and services.The advantages and disadvantages of the common classification features are analyzed for the common modulation modes of spatial signals: MPSK,MFSK,MQAM and MAPSK.The analysis shows that the high-order cumulant features cannot realize the MFSK signal classification,the wavelet transform features are not sensitive enough to MPSK,the classification effect is poor,and the single feature cannot realize the modulation recognition of the above four kinds of signals.In this paper,the combined features of high-order cumulant,entropy and wavelet transform are innovatively used,and the decision tree classifier is adopted to classify the above signals.Firstly,the recognition between categories is carried out,and the mixed signals are identified into three categories of MPSK,MFSK,MQAM and MAPSK by calculating the combined features of wavelet transform and entropy,and then the recognition of different orders within the categories is realized by constructing the functions of high-order cumulant features.Simulation results show that this method can classify four kinds of signals,and the recognition rate of mixed signals can reach 75% when the Eb/No is 5dB.