Encoding and decoding using constrained interleaving

摘要

Serially-concatenated codes are formed in accordance with the present invention using a constrained interleaver. The constrained interleaver cause the minimum distance of the serial concatenated code to increase above the minimum distance of the inner code alone by adding a constraint that forces some or all of the distance of the outer code onto the serially-concatenated code. This allows the serially-concatenated code to be jointly optimized in terms of both minimum distance and error coefficient to provide significant performance advantages. These performance advantages allow a noise margin target to be achieved using simpler component codes and a much shorter interleaver than was needed when using prior art codes such as Turbo codes. Decoders are also provided. Both encoding and decoding complexity can be lowered, and interleavers can be made much shorter, thereby shortening the block lengths needed in receiver elements such as equalizers and other decision-directed loops. Also, other advantages are provided such as the elimination of a error floor present in prior art serially-concatenated codes. That allows the present invention to achieve much higher performance at lower error rates such as are needed in optical communication systems.

主权项

1. A communications apparatus comprising:
a constrained interleaver configured to rearrange an ordering of a sequence of N=rm input bits in accordance with a permutation function in order to produce a permuted sequence of output bits, wherein N and m are positive integers, m>1, and the rearrangement of the ordering in accordance with the permutation function is equivalent to performing the following operations:
applying a word-level pseudo-random permutation function to pseudo-randomly rearrange an ordering of a plurality of multi-bit words embedded in the sequence of N=rm input bits;loading the N=rm number of input bits into a r×m array of bits, wherein the r×m array has r rows and m columns, and the bits are serially loaded into the r×m array of bits along rows,applying an ith pseudo-random row permutation function to each row i, for i=1, 2, . . . r, wherein the ith pseudo-random row permutation function pseudo-randomly rearranges an ordering of m bits in the ith row,reading bits out of the r×m array along r-bit columns to form the permuted sequence of output bits.