Classification And Recognition Of Communication Signals In Tactical Radio Network Based On Machine Learning Algorithm

Classification of the military communication signal's modulation mode is an important research hot in the signal processing field.The modulation mode needs to be recognized in the complex circumstance with noise and interference,if that is done,the next step of signal processing could carry on.Now,Tactical Radio Network is the battlefield communication network of America,it is based on the wireless communication,interconnects the tactical radio,information terminal and routing device,and it is the integrated battle and tactical communication system for US Army's digital battlefield.As the third party to investigate,the classification and recognition of the communication signal in the tactical radio network is very important in the electronic countermeasure.This thesis relies on the national project,mainly focusing on the classification and recognition of the communication signals in the Tactical Radio Network.The main work of the thesis can be summarized as follows:1.The parameters of the received Tactical Radio Network signal without prior knowledge have been estimated,including estimation of frequency hopping,symbol rate estimation.And the modulation classification of the Tactical Radio Network signals can be realized with this thesis' method that combined the instantaneous parameters,the parameters of the cyclic spectrum analysis and the wavelet packet decomposition and reconstruction parameters.The simulation results show that the support vector machine used as the classifica tion model has effectively realized the modulation classification of Tactical Radio Network signals.2.Based on the MATLAB simulation platform,some parameters in the characteristic parameter set mentioned above have been utilized to solve the problem of over learning which is existed in the support vector machine model.The support vector machine algorithm and radial basis function neural network algorithm have been respectively used to classify the modulation signals.Simulation results have showed that it can effectively realize the modulation identification of tactical radio network communication signals with the proposed method in this thesis.The performance of support vector machine model is better than radial basis function neural network model.Under high signal to noise ratio,the performances of them are both very good.But under low signal to noise ratio,the former still has brilliant recognition accuracy,the performance of the latter declined significantly.