Convolutional coupled codes between distance properties and convergence behavior of the iterative decoding scheme

The class of convolutional coupled codes is a promising alternative to classical turbo-codes. A convolutional coupled code consists of a cascade of ν identical recursive systematic convolutional (RSC) outer codes of rate k/n and k inner block codes with parameters (2ν, ν, d_i). The codes are linked together such that only the systematic part of the outer codes is encoded with the inner block encoders. The bits of each information vector of the outer codes are scrambled by a given interleaving before entering to the inner encoders. In contrast to parallel concatenated turbo-codes, in which the information symbols and the redundancy from the constituent codes are transmitted, we transmit only the redundancy produced by the outer and inner codes. The over-all code rate of the resulting code remains k/n. The influence of number, code memory and code polynomials of the outer convolutional codes on the distance properties and the convergence behavior of the iterative decoding scheme is studied.