A distance spectrum estimation method of turbo codes

Abstract

This study shows a new algorithm for distance spectrum (DS) estimation of turbo codes. We propose fundamental error event whose intervals of the consecutive two non zero inputs are less than the fundamental period. The DS is recursively estimated by some consecutive zero inputs with length of the fundamental period and the combination of fundamental error events. The proposed method also reduces the complexity of the calculation of codeword weight by looking up the tables of the number of weights and state transition instead of traversing over the trellis diagram.