Improved stopping criteria for iterative decoding of short-frame multi-component turbo codes

Abstract

In this paper, we investigate some techniques for improving existing stopping criteria for iterative decoding of short-frame turbo codes with two or more component codes. Conventional stopping criteria are applied only after every decoding iteration. We investigate the potential reduction on decoding complexity and delay by adapting conventional iteration-based stopping criteria to component-based variants with various step sizes, where the step size is a parameter we introduce for complexity minimization. If the step size is too small, the resultant stopping criterion becomes too easy to satisfy, causing performance degradation. If the step size is too large, the resultant stopping criterion becomes too difficult to satisfy, leading to unnecessarily high decoding complexity. It is demonstrated that for a frame size of 84 bits and E/sub b//N/sub o/=3 dB, impressive savings of 40%, 27% and 33% on the average number of iterations for iteratively decoding turbo codes with two, three and four component codes respectively can be obtained by improving the simple hard-decision-aided stopping criterion.