Parameter Estimation And Modulation Classification Signal In Stable Distribution Noise

Parameter estimation and modulation identification of communication signals is a key technology in non-cooperative communication system.It plays an important role in the field of electronic reconnaissanc e,radio spectrum monitoring,emergency rescue,and Internet of Things,etc.Modulation recognition refers to estimating the transmission signal modulation type and some parameter information directly from the received signal without knowing the specific i nformation of the modulation,providing the basis for signal demodulation,threat assessment,interference identification and other signal processings,etc.And the parameter estimation is the prerequisite for modulation identification,signal blind demodulation and other processing.At the same time as the application scene changes,the channel s often have some significant non-Gaussian impulse noise,the signal processing system designed by Gaussian model will be invalid.It is found that it is more effective to describe the impulse noise which is widely existed in life by stable distribution.Therefore,it is more practical to study the parameter estimation and modulation recognition of communication signals in the stable distributed noise environment.In this dissertation,the parameters estimation and modulation recognition algorithms are studied in Alpha stable distributed noise environment.(1)For the failure problem that parameters estimation of communication signal in the condition of stable distributed noise environment,the generalized cyclic spectrum is defined by constructing the nonlinear transformation method based on the traditional cyclic spectrum parameter estimation algorithm.The carrier frequency and symbol rate of the communication signal are estimated by using the position of the generalized cyclic spectral cross section line.BPSK signal is taken as an example to verify the algorithm.Finally,the BPSK and 16 QAM signals are verified by experiments.(2)For the problem that the high-order cumulant can not identify the communication signal in stable distributed noise environment,the generalized higher order cumulants are defined and the characteristic parameters are constructed by analyzing the different order cumulants of each signal.The characteristic parameters are used to identify the MFSK and MPSK signals.Then,the characteristics of the generalized cyclic spectral cross section lines of BPSK and QPSK,2FSK and 4FSK signals are analyzed.These features carry out i ntraclass recognition and design a classifier.Finally,the interclass and intraclass identification of 2FSK,4FSK,BPSK and QPSK signals are simulated by experiment.