Research And Implementation Of Automatic Digital Modulation Recognition

Automatic modulation recognition means that in a non-cooperative communication system,after processed and analyzed the received communication signal,we can determine the modulation method used by the signal,which provides a basis for signal demodulation and signal recognition.This technology has important applications in military and civilian fields,such as communication reconnaissance,electronic countermeasures,and signal supervision.The current automatic digital modulation recognition technology has the following challenges:1.With the increasing shortage of spectrum resources,high-order modulation methods with higher resource utilization have been widely used.In addition,the complex electromagnetic environment also makes blind recognition under low SNR a real challenge.Therefore,it is necessary to propose a new blind recognition algorithm to improve the accuracy of blind recognition;2.The traditional blind recognition is performed by professional operators through the observation of the spectrum analyzer which shows the signal picture,to determine the modulation method used by the signal.Therefore,it is necessary to design and implement a blind recognition platform to replace manual operation and improve the recognition efficiency and accuracy of blind recognition.The research content of this article is as follows:1.Aiming at the problem of the modulated signals cannot be well characterized by the limited features,an automatic modulation recognition algorithm based on joint features is proposed.This algorithm firstly completes the feature extraction of the signal by means of clustering,and extracts the two features of the membership matrix and the location of the cluster center;secondly,the feature vector is generated based on the location of the cluster center,and combined with the existing feature vector based on the membership matrix,which expands the feature vector that characterizes the signal;finally,the feature vector is input to the designed shallow neural network classifier,and output the signal's modulation method.The simulation results show that when the signal-to-noise ratio is 0dB?4dB,the overall recognition rate of the automatic modulation recognition algorithm based on joint features is increased by 20%-30%compared with the comparison algorithm based on single feature.When the ratio is 0dB,the blind recognition accuracy of 64QAM high-order modulation is improved by 36%compared with the comparison algorithm based on a single feature.2.Aiming at the problems of the automatic modulation recognition platform cannot be accessed across devices,the realization complexity is high,and the recognition accuracy rate is low,the automatic modulation recognition platform is designed and implemented.First,the platform needs analysis and overall design,and the functional modules and technical solutions of the platform are determined;secondly,five functional modules are designed and implemented,namely,data input module,data preprocessing module,automatic modulation recognition algorithm module,and data analysis and calculation module and data visualization module,these modules are used to realize and optimize the function and performance of the automatic modulation recognition platform.Finally,the function verification and performance analysis of the platform are carried out.The results show that the platform can be accessed across devices,with low implementation complexity and high recognition accuracy.