Voice Code Type Detection Of Digital Walkie-talkies Based On Two-dimensional Signal Recombination And Fusion Filtering

With the rapid development of science and technology,modern society has entered into the information society and digital society.In the field of communication,the digital walkie talkie is also gradually replacing the analog walkie talkie.Compared with the traditional analog walkie talkie,the digital walkie talkie has the advantages of high utilization rate of spectrum resources,high call quality,strong security,strong anti-interference ability.However,with the rapid development of digital walkie talkies,it has been applied to many negative situations.In recent years,the use of digital interphone cheating in examinations over the country frequently has brought great social impact.The vocoder signal detection system with digital walkie talkie is an effective method to prevent cheating in examinations.It uses the detection receiving device to monitor the radio environment around the examination room in real time and collect the digital walkie talkie vocoder signal.The characteristics of digital walkie talkie vocoder signal are analyzed to determine whether the collected signal is a cheating signal.In order to solve the difficult problem of digital walkie talkie vocoder signal detection,we adopts support vector machine and kernel extreme learning machine as the classification framework.Based on analyzing and extracting the features of digital walkie talkie vocoder signal,we propose a digital walkie talkie vocoder signal detection method based on the fusion filtering on two-dimensional recombination signal.Then an enterprise-level digital walkie talkie vocoder signal detection system is designed and implemented.Experiments on real digital walkie talkie vocoder signal data verify the validity of the detection methods and system.The main works of the papers include:(1)Experiments verify the validity of support vector machine and kernel extreme learning machine on digital walkie talkie vocoder signal detection.Based on analyzing the features of digital walkie talkie vocoder signal,we use principal component analysis and propose a digital walkie talkie vocoder signal detection model based on support vector machine and kernel extreme learning machine.Experiments based on the real digital walkie talkie vocoder signal data show that support vector machine and kernel extreme learning machine can be effectively applied to the digital walkie talkie vocoder signal detection system.(2)A digital walkie talkie vocoder signal detection method based on the fusion filtering on two-dimensional recombination signal is proposed.In order to overcome the disadvantage of one-dimensional feature,we apply fusion filter on two-dimensional signal which is recombined from original one-dimensional signal.Then,a variety of signal feature types are selected for analysis according to the characteristics of the digital walkie talkie vocoder signal.The attribute reduction method is used to select the most effective features from the feature set which is constituted by the extracted features,which compress the high-dimensional feature space to the low-dimensional feature space to achieve high-precision detection of vocoder signals.Experiments show that the proposed method can achieve high-precision detection on digital walkie talkie vocoder signal.(3)Based on above methods,we design and realize an enterprise-level digital walkie talkie vocoder signal detection system based on MFC framework.The system consists of three modules,including file management and display,model training module,vocoder signal detection module and son on.We provide the system framework,main process design and software core module development and testing.