| 摘要 |
The present invention relates to an improved decoding method capable of solving a problem of an error floor of a correction code from a family of a turbo code, a low density parity check (LDPC), or turbo-type codes and maintaining the same spectrum efficiency, namely, without any reduction in useful throughput of an encoded stream. This result is obtained by identifying bits significantly affected by an error, when input to a decoder, and changing likelihood corresponding to the bits, thus improving convergence of the decoder. |
