TY  - GEN
T1  - An Efficient Turbo decoding over Impulse Noise Channels
A1  - Yunghsiang S. Han
A2  - Yunghsiang S. Han
LA  - en_US
YR  - 2014
UL  - http://buscaintegrada.pucsp.br/vufind/Record/NDLTD-TW-102NTUS5442017
AB  - 博士 === 國立臺灣科技大學 === 電機工程系 === 102 === Impulse noise is a serious problem in communications. This intense noise
hinders transmitted signal and raises the error of exchanged data. Some previous
works on channel coding adopt traditional codes and signal processing to solve this
issue, but encountered many difficulties in handling the impulse.
Regarding impulse noise, first we analyzed the impact of impulse noise to turbo
code, which is a robust and popular code in white Gaussian channel. Referring to
metric-erasure Viterbi algorithm, we developed a metric clipping turbo code based on
partial information of impulse model. To evaluate this code in non-Gaussian channel
exactly, we introduced kernel density estimation into extrinsic information transfer
band chart for operating evaluation.
The result of computer simulation and evaluation shows the performance is close
to model-modified turbo code and the complexity is almost the same as primary turbo
code.

ER  -