Abstract:

In order to avoid fails of blind parameters identification for some particular low code rate linear block codes and improve the identification performance of common linear block codes， a blind identification approach based on binary column Gaussian elimination (BCGE) is proposed. Firstly, an intercepted sequence is put into an analysis matrix horizontally, then the rate of “1” involved in each column of the matrix after BCGE is computed. Finally the mean and variance of all the rates of “1” are calculated. In good conditions of synchronization with a very low error rate, the code length is estimated with means difference， in poor conditions of asynchronization with a high error rate, the code length is estimated with variances difference. Theoretical analysis and simulation shows the blind identification approach has a good faulttolerant performance and can identify the code length and synchronization of linear block code of various code rates in low computation complexity.