Improved method for searching interleavers from a certain set using Garello's method with applications for the LTE standard

摘要

In this paper, we propose a method for searching interleavers within a certain class, with the aim of designing turbo codes with good distance spectrum. The method is based on a modified version of Garello's algorithm and consists in the calculation of frame error rate truncated upper bound. Here, it is applied to quadratic permutation polynomial (QPP) interleavers able to outperform those chosen for the long-term evolution (LTE) standard, for lengths up to 1,504 bits. Three classes of interleavers have been analyzed: (1) the set of QPP interleavers with the largest spread, (2) the set of QPP interleavers with a spread parameter equal to that of LTE interleaver and the highest refined nonlinearity degree, and (3) the complete set of all QPP interleavers for lengths up to 1,008. The distance spectrum optimization is made for all classes. Compared to previous methods for finding QPP-based interleavers, the search complexity is reduced, with improved performances in terms of search time, allowing interleavers of higher length. For lengths up to approximately 450, the best interleavers were found in the first class. For longer lengths, the second class contained the best ones.