Digital Modulated Signals Modulation Recognition And Parameters Estimation

With strong ability of anti-interference, easy to store and process, easily encrypt and easily integrated, digital modulated signals are widely used in various kinds of communication fields. Along with the increase in signal types and modulation complexity, automatic and efficient identification of signal modulation type and estimation of some parameters have become the research hotspot. According to the characteristics of the communication signals in spectrum monitoring field, the modulation recognition and parameters estimation problems are studied in this paper.The existing signal modulation mode recognition methods can be roughly divided into two broad categories:likelihood ratio decision theory category and signal feature extraction category, each category contains many different methods and have its advantages and disadvantages, so we need to select the appropriate method according to different problems. Signal parameters estimation is mainly achieved through the methods of extracting signals’ characteristics. This part will be described in the first chapter of this paper.Although the research of signal modulation recognition and parameters estimation has been studied for a long time, there are still some problems at present. Especially in spectrum monitoring area, the received signals have none prior knowledge, a wide variety of signals to be identified, the communications environment is becoming even worse and the difficulty to extract stable features make modulation recognition and parameters estimation very complicated in practical processing.As it is difficult to extract the signal stability characteristics, the paper proposes a new method based on wavelet packet transform and the SVD transform to recognize some intra-class digital modulated signals. Experiments prove that the extracted features under the condition of different carrier frequency and symbol rate are relatively stable. After the intra-class recognition, the next work is to distinguish inner-class signals. According to the characteristics of different modulation types, this paper respectively selects the cyclic spectral density features, the instantaneous frequency features and constellation diagram features to realize the recognition of inner-class signals of MPSK, MFSK and MQAM modulated types, then using computer to perform simulation and analyzing the results. Another work of the paper is the signal parameters estimation. Square spectrum and power spectrum method are used to estimate signals" carrier frequency, and the estimated performance differences are discussed. Using the method of wavelet coefficient spectrum feature for symbol rate estimation, experiments prove that the method has very good effect; the method of using the inflecting point of signal power spectrum distribution function is used to calculate signal-to-noise ratio, experiment show that the method is almost not influenced by signal modulation type.Considering that the a stable distribution could reflect the characteristics of the noise more effectively relative to the Gaussian distribution, the paper proposes an improved second-order statistics method to estimate some parameters of digital modulated signals under a stable distribution. The improved cyclic spectral density function has the ability of limiting the signals’amplitude but doesn’t changing the signal phase information at the same time, so as to ensure the improved cyclic spectral density function has the same characteristics with the original one. Then we use parts of slices in cycle frequency field as the features to estimate the carrier frequency and symbol rate. Theoretical derivation and simulation results show that the method proposed in this paper can greatly enhance the performance of parameters estimation and has certain practical value.