Network-on-chip-centric approach to interleaving in high throughput channel decoders

Reliable wireless communication needs efficient channel coding schemes like turbo and LDPC codes. During decoding, data is exchanged iteratively between component decoders. Between iterations, however, the data blocks are subjected to a permutation (or interleaving). As parallelization of the decoder architectures is mandatory for high throughput, access conflicts can occur. For standard compliant decoders it is impossible to design, or pre-process, the permutation patterns such that conflicts are avoided. Therefore, we employ networks-on-chip capable of resolving access conflicts at run-time to support arbitrary interleavers without any pre-processing.