Research And Implementation Of Modulation Recognition Algorithm For Distributed Electromagnetic Sensor Network

So as to realize the modulation recognition of the wireless communication signal and emitter location,Electromagnetic spectrum monitoring through miniaturization and low power distributed wireless sensor networks is an important way to effective supervision of radio spectrum resources.In this paper,we focus on the feature extraction,classification and recognition algorithm design,and distributed information fusion processing method for the wireless sensor network node energy constraints and the diversity of the monitoring signal set.Finally,the distributed electromagnetic spectrum monitoring simulation training system is implemented.The specific research contents are as follows:1.According to diversity of monitoring signal monitoring system in the electromagnetic spectrum,this paper proposes a multi feature combination algorithm for modulation recognition based on support vector machine,it realizes the effective identification of monitoring signals of seven kinds of concentrated signal by using only three features.The first design of high order cumulants recognition algorithm based on support vector machine,it realizes the classification and recognition of MASK,MPSK,MFSK,MQAM and other ten kinds of signals,the simulation results show that the SVM recognition algorithm is better than the decision tree.On this basis,this paper puts forward signal recognition by high-order cumulant and cyclic spectrum feature,it completes the recognition of 2ASK,BPSK,QPSK,8PSK,2FSK,MSK,16 QAM seven kinds of monitoring signals through minimal characteristics of signal.The simulation results show that when the signal to noise ratio(SNR)is 5d B,the recognition rate of the seven signals has reached more than 80%.2.For the multi node identification algorithm in the lower SNR situation,the performance of the proposed algorithm is not satisfactory,a distributed modulation recognition algorithm based on D_S evidence theory is proposed.Based on the D_S evidence theory,a decision level fusion algorithm based on two levels of feature element and feature vector is proposed.In this paper,a new method is proposed to calculate the basic probability distribution function of the attribute element level by calculating the Mahalanobis distance of each attribute element of the feature vector and each signal.The basic probability distribution function of element level is fused in the local node,and the basic probability distribution function of each node is obtained.Finally,the fusion result is obtained at the center node.The simulation results show that the performance of the proposed algorithm is better than that of the original multi node cooperative identification algorithm at low SNR.Finally,the fusion recognition algorithm is used to identify the signal set by the joint identification.When the signal to noise ratio(SNR)is 2d B,the recognition rate of the seven signals has reached more than 75%.3.Key techniques of distributed electromagnetic spectrum monitoring simulation training system is implemented.Firstly,the overall framework and software design and implementation of the distributed electromagnetic spectrum monitoring simulation training system are introduced.The monitoring module mainly realizes the function of the detection of radiation source signal,waveform display,signal recognition and the extraction of important parameters.At the same time,the proposed algorithm is embedded into the monitoring system,and the system level algorithm performance is tested.The results are similar to the theoretical analysis,which shows the overall correctness of the recognition algorithm.