Study On Spectrum Intelligent Sensing And Interference Intensity Estimation In Measurement And Control System

In the space TT&C system,the channel environment between the TT&C equipment and the spacecraft is unstable,and the interference signals received by the TT&C equipment(signals sent by other equipment are interference signals to the TT&C equipment)are greatly affected by noise,with a large dry noise ratio interval,and in a large interval,the dry noise is relatively low.In measurement and control equipment,need to do to noise ratio in large dynamic range,dry ratio range of low s/n ratio under the condition of dry docking received signals to judge whether there is other users,the interference signal is received dry ratio estimation,received signal modulation mode recognition,to complete measurement and control device of received signal demodulation,signal modulation mode selection and use of frequency band selection.Therefore,this paper is divided into three parts: spectrum sensing technology,interference signal intensity estimation technology and modulation pattern recognition technology.In recent years,researchers have made great breakthroughs in the study of neural network,which has been widely used in image and voice,etc.Neural network has a good ability to automatically extract deep robust features,so this paper USES neural network to study these three contents.First,Study on spectrum intelligent sensing technology based on neural network.In the research process,the general spectrum sensing algorithm adaptive energy algorithm(ADAPT-ED)and eigenvalue detection(MME)were used for spectrum sensing.In order to ensure that the spectrum sensing method can be in a larger signal-to-noise ratio range and lower signal-to-noise It has better performance than the following.First,the single-channel neural network spectrum sensing technology is proposed.The single-channel spectrum sensing technology only considers the characteristics of time domain information,and does not consider the characteristics of frequency domain information.Then it further proposes a dual-channel neural network.Spectrum sensing technology.Simulation experiment results show that single-channel spectrum sensing technology can achieve 90% detection probability under the condition of dry noise ratio of-12 d B and above than traditional method of spectrum sensing technology.Dual-channel neural network spectrum sensing technology is better than single-channel neural network spectrum sensing The technology has better performance under the condition of low interference ratio,and the detection probability reaches 90%under the condition of interference ratio of-15 d B and above.Secondly,study on interference Signal strength estimation technology.Thetraditional dry noise estimation algorithms include the one based on autocorrelation matrix eigenvalues and the one based on M2M4.Through the study of the two traditional algorithms,it is found that the estimation performance of the traditional algorithm in the larger dry noise ratio interval is poor,and the estimation performance of the interference signals with different modulation modes is greatly different.Therefore,the general neural network interference signal strength estimation method is first proposed.Due to the advantages and disadvantages of convolutional neural network and cyclic neural network in data processing,the feature fusion based interference signal strength estimation technique is further proposed.The simulation results show that the performance of the neural network based interference signal strength estimation technique is greatly improved compared with the traditional algorithm,and the performance of the feature fusion based interference signal strength estimation method is better than that of the general neural network.Finally,study on interference signal modulation pattern recognition technology.Because the modulation methods for different interferences are different,the range of the interference to noise ratio is large,and the interference to noise is relatively low.It is difficult to recognize the modulation patterns of signals with a variety of modulation methods in the large dynamic interference to noise ratio range.The neural network has a good automatic extraction.Stick features and good generalization ability,so the interference signal modulation pattern recognition technology based on neural network is proposed.Noise interference has a certain effect on the feature extraction of interference signal,so the modulation pattern recognition technology based on noise reduction neural network is further proposed.The simulation experiment results show that the modulation pattern recognition technology based on neural network can achieve80% recognition accuracy for most modulated signals in the interference-noise ratio environment above 0db.After the noise reduction processing of the GRU noise reduction network,it can achieve 0d B The above recognition rate for most signals is90%,and the recognition performance at a lower interference-to-noise ratio improves the signal performance for different modulation methods.