An Improved Blind Recognition Method of the Convolutional Interleaver Parameters in a Noisy Channel

In a digital communication system, forward error correction (FEC) codes and interleavers are implemented to code the data so as to improve the error performance, which is hindered by random and burst errors. In the context of noncooperative communication, interleaver parameter recognition, which is the prerequisite for frame synchronization, channel coding recognition subsequently, is of vital significance. Methods of blindly recognizing convolutional interleaver parameters have been proposed in published literature, but the accuracy of recognition decreases significantly when bit error rate (BER) is high. To improve the performance of the algorithm, the effect of error bits on Gauss-Jordan elimination through pivoting (GJTEP) algorithm is analyzed in this paper. The following conclusion is drawn: error bits on the principal diagonal of data storage matrix will exert a great impact on the recognition accuracy. Based on the conclusion, an improved blind recognition method with denoising, the core principle of which is reducing error bits on the principal diagonal of data storage matrix, is proposed in this paper. The simulation experiment results demonstrate that the performance on error resilience is markedly improved.