Spatialt kopplade turboliknande koder
2014-01-01 -> 2017-12-31
Communication over a physical channel suffers from errors that occur during the transmission and channel coding is a technique that allows to correct them. The invention of turbo codes in the early 90's and the rediscovery of low-density parity-check (LDPC) codes have led to a breakthrough in coding theory: Shannon's channel capacity, the ultimate data rate limit for reliable transmission, could finally be approached. But this performance is achieved only for very long blocks while for shorter blocks, which are of particular importance in practice, turbo codes and LDPC codes are penalized by a flattening of the error rate curve.
The purpose of this project is to investigate, understand, and design codes on graphs with efficient decoding algorithms for powerful error correction in communication systems. In particular, this project aims at addressing the fundamental open issue in coding of achieving both close-to-capacity performance and very low error rates for block lengths of practical interest, by applying the recently introduced concept of spatial coupling to turbo-like code constructions. For shorter block lengths we expect these codes to achieve better error correction capability than spatially coupled LDPC codes.
Our research falls into the category of fundamental research. However, the proposed research is also very relevant for reliability and latency constrained applications like machine-type communications and vehicular-to-vehicular communications.
Alexandre Graell i Amat (Chalmers)