Research And Implementation Of Digital Signal Modulation Recognition In Wireless Communication Environment

Modulation recognition technology is one of the important non-cooperative communication technologies such as spectrum detection,threat coefficient evaluation,electronic countermeasures,etc.It is widely used in both civil and military fields.In the existing modulation recognition methods,most of them have problems such as few recognition types,poor recognition performance of high-order modulation signals,and low recognition rate under low SNR.The research on digital signal modulation recognition technology is launched to address these issues in this paper.The main work results of the paper are as follows:Systematically analyze the modulation principles of common digital signals,including:MASK,MPSK,MFSK and MQAM,and experiment and analyze the modulation process and characteristics.Three traditional modulation recognition methods based on feature parameters extraction of instantaneous parameters(IPs),Wavelet transform(WT),and high-order cumulants(HOCs)are introduced.The advantages and disadvantages of each methods are summarized,and some methods are experimentally reproduced.The decision tree classifier is introduced,and its advantages and disadvantages in the process of modulation recognition are analyzed and summarized.Aiming at the problems of the traditional method of IPs with fewer recognition types,low recognition rate of high-order modulation signals and frequency modulation signals,a combined modulation recognition method of power spectrum entropy(PSE)and IPs is proposed.The improved feature parameters of amplitude and phase are selected to identify amplitude and phase modulation signals.The feature parameter of PSE is introduced to identify the MFSK signal within the class.After obtaining a suitable threshold,the method can effectively identify 9 digital modulation signals when the SNR is greater than 6dB.Compared with the related literature can find that the method in this paper is superior to the traditional method and the improved method in terms of identifying the modulation type and the recognition rate under low SNR.Aiming at the disadvantages of most existing methods of modulation recognition like susceptibility to noise interference,low recognition rate under low SNR,and the need to combine multiple feature parameters,a method of modulation recognition with feature parameters constructed only by HOCs is designed.Starting from the properties of HOCs,the HOCs of common digital signals are deduced,and the first-order differentiation and median filtering are performed on MFSK signals,which solves the problem that HOCs don't contain frequency information.Constructed a number of characteristic parameters composed of fourth-order and above HOCs.After simulations have verified the correctness of the calculated theoretical values,appropriate identification rules are designed to achieve effective identification of 10 modulation signals.Simulation experiments and analysis were made on the reasons for the low recognition rate of 8PSK and others under low SNR and the influence of the length of the signal to be measured on the recognition rate.Compared with the literature of related methods can find that the method of paper has a greater improvement in recognition rate and recognition types than the related method of HOCs and the combination method of HOCs and WT,indicating that the method has good universality and superiority.