Research On Automatic Recognition And Parameter Estimation Of Digital Communication Signals

Various modulation techniques are widely used in the wireless communication systems which occupy various frequencies and bandwidths. With the rapid development of communication systems, we have to face the problem of contention and allocation of the limited frequency resource. The automatic recognition and parameter estimation techniques of communication signals can be utilized in the authentication and spectrum management for radio signals. As the important ingredient of electronic supports, the automatic recognition and parameter estimation techniques provide the information which is the foundation of the electronic attack and protection. At the same time, the techniques are the inherent requests of the software radio development. Consequently, it is of great actual value to study the automatic recognition and parameter estimation techniques for communication signals.Based on the previous works, this paper primarily investigates the automatic classification of modulation format and the parameter estimation techniques for digital communication signals associating actual projections. The main works can be summarized as follows:1. A novel algorithm is provided, which can effectively estimate the symbol rate of the digital signals using the Wavelet Transform (WT) and the spectrum analysis. The power spectrum of the WT coefficient modulus has discrete spectral lines located at the integer-multiplied symbol rate, so that those spectral lines can be used to detect the symbol rate of the digital baseband signals. This algorithm is simple and has better estimated accuracy. Different digital baseband signal corresponds to different density function, which makes signal classification possible. The probability density function estimation and support vector machines classifier is used to classify the digital baseband signals automatically. The advantages of this algorithm lie in the fact that it is easy to construct training samples in the condition of small sample study, which meets the application requests of the signal detection. Better recognition results can be yielded for many digital baseband signals in various noise background.2. Under the software radio frame for the modulated signals'detection, the corresponding processing algorithms of digital signals are adopted to abstract the classifying features for the MASK, MFSK, and MPSK signals. Thus the algorithm adopting the density function estimation and the support vector machines classifier...