Research On Digital Modulation Recognition

Digital modulation recognition is a hot research field in wireless communication,especially compared to demodulation,decoding and other technologies.Digital modulation recognition can also play a significant role in communication.However,some complex changessuch as noise interference and signal-to-noise ratio change bring great challenges to modulation recognition.Digital modulation recognition algorithms still need to make more in-depth research.In order to obtain a digital modulation recognition algorithm with a higher recognition rate,based on a full understanding of the background and existing algorithms,as to the used widely digital modulation signals,the research work is completed as follows:(1)A digital modulation recognition algorithm based on neural network is researched.The overall framework and flow of the proposed algorithmis given.Firstly,convert the received signal,extract the characteristic parameters,and then send the signal to a mature neural network classifier that first obtains a rough identification target through the training network for accurate detection.Finally,highlyrobust modulation signal recognitionis obtained.The simulation experiment results show that:Using this algorithm,six kinds of digital modulation signals including 2ASK,4ASK,2FSK,4FSK,BPSK,and QPSK are successfully classified and recognized on the MATLAB platform,and the recognition rate is stable at more than 90%.(2)A digital modulation recognition algorithm based on higher-order cumulants is researched.A comprehensive analysis and discussion of the higher-order cumulants of the modulation signal are made.The modulation recognitionmethodof seven kinds of digital signals: BPSK,QPSK,8PSK,8QAM,16 QAM,32QAM and 64 QAMare discussed.Firstly,extract the feature parameters according to the algorithm requirements.Then use MATLAB to perform relevant simulation experiments.Finally,observe the recognition success rate of the algorithm under different signal-to-noise ratios.According to the experimental results,it can be seen that the algorithm can meet the basic recognition requirements of modulated signals.(3)Since the constellation diagram can describe the spatial distribution of the signal,a digital modulation recognition algorithm based on constellation clustering is designed.Aiming at the digital modulation signals of MQAM and MPSK,an improved feature recognition methodis proposed.Firstly,based on empirical clustering baselines,usesubtraction clustering to obtain more reasonable cluster centers and get the number of clusters.Then,based on the number of cluster points and cluster positions,a new cluster center based on the number of cluster centers and the circle scale ratio of the constellationis defined.Finally,MQAM signals can be recognized successfully.The feature recognition function considers that the MPSK signal constellationhas the same circle radius,and redefines the feature function to achieve the modulation recognition of MPSK signals according to the distance from the clustering point(excluding the point on the y-axis)to the x-axis to the origin point from the minimum phase of the y-axis.Simulation results show that the improved method replaces the traditional subtractive clustering method,and the recognition rate is higher.