The Application Of Modulation Recognition Algorithms In The Radio Signal Monitoring

Radio signal monitoring refers to the detection,search and receive modulated radio signals,and analysis,identify and monitor the feature parameters of signals.Identifying and estimating the feature parameters of the modulated signals without any priori information is the first stage of the signal monitoring.Hence,the recognition of the modulated signals plays an important role in the field of radio signal monitoring.This paper tries to provide a general strategy to the modulation recognition by combining the existing advanced methods.Specifically,we have the following contributions.1)We introduce 12 types of modulated signals with rigid mathematical model,namely,amplitude modulated signal,frequency modulated signal and phase modulated signal.Then we discuss the features of the corresponding modulated signals from the instant amplitude,instant frequency and instant phase.We also discuss the impact of noises to the identification of the modulated signals.2)We try to identify all types of the modulated signals with a multi-steps strategy by combining the advanced algorithms,e.g.,Kernel fisher discriminate algorithm,wavelet transform method and modulus maximum modulation signal feature extraction algorithm.We classify the modulated signals with a combination method,and ultimately achieve the overall recognition of all kinds of modulated signals.3)Specifically,we introduce the existing algorithms for the modulated signals recognition,and we propose our general method by a three-stage scheme: first,we introduce the maximum spectral density feature method to classify 12 types of modulation signal as two categories: the amplitude signals and the frequency modulation,phase modulation signals.As the amplitude modulated signal is non-constant envelope signal while the frequency and phase modulation signals are constant envelope signals,such characteristic can be used to distinguish these two categories.Secondly,we employ the modulus maximum modulation signal feature extraction method to classify the analog amplitude modulation signal and digital modulation signal.Moreover,the Kernel fisher discriminate algorithm is used on the discrimination between the simulated signals and the digital signals.During this stage,we also employ the wavelet transform method as the preprocessing of de-noising for the modulated signals.In the third stage of modulation recognition,wavelet analysisand kernel discrimination analysis are used to identify each specific modulation types of the classified signals.Comparing the proposed algorithm with the modulated signal identification algorithms through experimental simulation results,the proposed scheme can significantly improves the robustness of identification.For the simulation experiments part,the results illustrate that the multi-layer identification algorithm can ensure above 90% precision under the case of low SNR noises.