Study On Short Data Processing In Communication Blind Receiver

In communication blind receiver, it's of great significance to study how to estimate signal's parameters and achieve blind equalization form short data series. Based on field project, this dissertation focus on the symbol rate estimation and blind equalization techniques from short data series (from several decades to several hundreds symbols). The main work and the results are summarized as follows:Firstly, aiming at the symbol rate estimation method by second nonlinear transform showing poor performance in the case of short data and less spectrum redundancy, a modified algorithm was introduced: Filtering the received series first by a high pass filter, then applying second nonlinear transform. The analyses shows that the modified algorithm improved the performance of the original algorithm greatly in the case of short data and less spectrum redundancy. Simulation also verified the conclusion. K.C.Ho's symbol rate estimation method by wavelet transforms show some defects and instability, the reason which induced these defects was analyzed in this dissertation, and then a modified algorithm was introduced: Removing the residual carrier frequency, and then applying wavelet transform. This method get rid of the defects of the original method thoroughly, improved the performance of the wavelet method in the case of short data series. Simulation verified the conclusion.Secondly, aiming at the difficulty of blind equalization by short data in length of several decades to hundreds, the short data series reusing method in Bussgang algorithm was analyzed in this dissertation. First of all, the reason why the short data series reusing method to be effective was analyzed, and then the facts which will influent the effect of reusing short data series in Bussgang algorithm was analyzed. By these analyzing results, some conclusions were got about how many data the Bussgang algorithm needed to achieve equalization: By reusing short data series, some Bussgang algorithms can achieve equalization by only several decades'data, which is close to the L.Tong's method. Thirdly, blind receiver put many requirements to blind equalization method, such as less computing consuming, less data consuming, and common adaptability to all kinds of communication signal. A new equalization method was designed in this dissertation to meet this requirement. First of all, the essence of Bussgang algorithm's convergence was studied, and the conclusion was got that Bussgang algorithm is one kind of the Minimum Entropy algorithm. By this conclusion the common cost function of Bussgang algorithm was put forward. This common cost function also indicates a new way to design new Bussgang algorithm. Some new Bussgang algorithms was designed according to this method, simulation has verified the convergence of these new algorithms. Combining the merits of these new algorithms, a dual-mode blind equalization algorithm was designed. This dual-mode algorithm shows some excellent performance: less computing consuming, less data consuming, and great adaptability to many kind of communication signals, which makes it meet to the requirement of blind receiver well. This new dual-mode algorithm has been used in communication blind receiver.Finally, based on software radio techniques, a communication blind receiver was designed, which integrated the algorithms designed by this dissertation.