SearcharxivSearch

arXiv subjects

Valentin Savin

Publications and source records attributed to Valentin Savin.

39 records · Page 3Linked to original sources

Minimum-Delay Decoding of Turbo-Codes for Upper-Layer FEC

In this paper we investigate the decoding of parallel turbo codes over the binary erasure channel suited for upper-layer error correction. The proposed algorithm performs on-the-fly decoding, i.e. it starts decoding as soon as the first symbols are received. This algorithm compares with the iterative decoding of codes defined on graphs, in that it propagates in the trellises of the turbo code by removing transitions in the same way edges are removed in a bipartite graph under message-passing decoding. Performance comparison with LDPC codes for different coding rates is shown.

cs.IT

Algebraic-geometric codes from vector bundles and their decoding

Algebraic-geometric codes can be constructed by evaluating a certain set of functions on a set of distinct rational points of an algebraic curve. The set of functions that are evaluated is the linear space of a given divisor or, equivalently, the set of section of a given line bundle. Using arbitrary rank vector bundles on algebraic curves, we propose a natural generalization of the above construction. Our codes can also be seen as interleaved versions of classical algebraic-geometric codes. We show that the algorithm of Brown, Minder and Shokrollahi can be extended to this new class of codes and it corrects any number of errors up to $t^{*} - g/2$, where $t^{*}$ is the designed correction capacity of the code and $g$ is the curve genus.

cs.IT

Iterative LDPC decoding using neighborhood reliabilities

In this paper we study the impact of the processing order of nodes of a bipartite graph, on the performance of an iterative message-passing decoding. To this end, we introduce the concept of neighborhood reliabilities of graph's nodes. Nodes reliabilities are calculated at each iteration and then are used to obtain a processing order within a serial or serial/parallel scheduling. The basic idea is that by processing first the most reliable data, the decoder is reinforced before processing the less reliable one. Using neighborhood reliabilities, the Min-Sum decoder of LDPC codes approaches the performance of the Sum-Product decoder.

cs.IT