The Research On Fm Broadcast Signalrecognition Based On Clustering Analysis

With the rapid development of the national economy and national defense construction,radio spectrum resource has become increasingly scarce.As an important application of radio technology,Frequency Modulation(FM)Broadcast plays an important role in enriching lives,ensuring traffic safety,disaster rescue and so on.In order to solve interference and illegal use,a new recognition method of FM Broadcast signal based on clustering analysis is proposed in this paper.Different from Auto Speech Recognize(ASR)of FM Broadcast,in terms of frequency domain,detail information including bandwidth,signal strength and waveform is obtained to extract feature.Then clustering analysis based on density is used to recognize FM broadcast.Furthermore,it is applied to discover illegal broadcast.Considering excellent efficiency of clustering analysis itself,experiment shows that our method is more suitable than traditional graph matching algorithm.The core work of this paper is as follows:1.Data preprocessing and feature extraction: According to obvious difference of every frequency,extract six feature——mean,variance,short-time average energy,normalized peak,kurtosis,zero center normalized standard deviation.Thus,a six-dimension feature space is constructed to reduce dimension.The experiment shows these six feature are extremely representative for real-time data.2.FM Broadcast signal recognition based on clustering analysis: In this paper,we describe core idea of the clustering based on peak density,Standardized Euclidean Distance is to improve the above mentioned clustering analysis and achieve good results.A large number of experiments show that the proposed improved method is feasible to identify mute signal,voice signal and music signal.3.Discovery of illegal broadcast: Collect IF Measurement data of integral point,extract the above six feature,calculate Standardized Euclidean Distance,identify FM Broadcast signal based on the improved clustering analysis.Finally,cluster centers of legal broadcast and illegal broadcast are given and applied to identify illegal FM Broadcast signal.Experimental results show that our method is efficient to identify illegal broadcast and legal broadcast.