Blind detection of interleaver parameters for non-binary coded sequence

Abstract

Recently, inter-leaver has become an indispensible component in communication systems. It permutes the sequence of data and hence efficiently reduces the effects of fading, especially in the form of burst errors. In a non-corporate context, blind estimation of inter-leaver parameters is extraordinarily important during the de-interleaving process. In this project, we endeavor to evaluate and investigate the schemes of recovering inter-leaver parameters for both binary and non-binary data sequences. The inter-leaver parameters can be easily retrieved in a perfect channel. The inter-leaver size is obtained based on the rank criterion. However, there is no simple formula to solve this problem when additive noise makes the deficient matrix full-rank. We attempt to analyze the Gaussian eliminated matrix to differentiate dependent columns from independent columns. For binary data, we count the number of ones in each column of a rank-deficient matrix. Threshold is set to categorize dependent and independent columns. For non-binary data, the rank-deficient matrix is classified by its large mean and variance of the percentage of zeros of a column. The performance of this algorithm is evaluated in both Gaussian and fading channels. Based on a thorough analysis of existing models, we put more effort forth for achieving automatic detection of the inter-leaver size for non-binary data and improving the performance of the algorithm within the low SNR (signal to noise ratio) region. We implement SVD (singular value decomposition) to recover the rank of a matrix in the presence of noise. Notably, the rank is also beneficial to perform a blind synchronization of the inter-leaved blocks. Moreover, the lower limit of SNR for the satisfactory performance is significantly dropped by reordering the original matrix which scales down the vulnerability of the Gaussian elimination algorithm under the attack of noise.