Robust Decoding Schemes for Coded Transmission Through a Markov Gaussian Channel

• Main Author: Fikreselam Gared Mengistu
• Other Authors: Yunghsiang S. Han
• Format: Others
• Language: en_US
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/40741379520610580472

博士 === 國立臺灣科技大學 === 電機工程系 === 102 === Communication systems are susceptible to impulse noise, particularly when the impulse statistics are not time-invariant and are difficult to accurately model. To address the challenge of impulse noise, a robust and efficient Viterbi and turbo decoding schemes were devised over memory impulse noise channels. By accommodating channel states, but without relying on statistical knowledge of impulses, the Viterbi algorithm (VA) based on an expanded set of trellis states, was employed to perform maximum likelihood decoding. A detailed analysis of complexity for the proposed Viterbi decoding was offered; the analytical results reinforced the efficiency of the proposed scheme compared with the traditional VA. We also proposed robust turbo decoding algorithm over a Markov Gaussian channel. The branch metrics obtained from the two- dimensional trellis diagram of the proposed VA was adapted to propose robust turbo decoding scheme. The simulation results indicated that the proposed Viterbi and turbo decoding schemes are compellingly robust: the bit error probability performance level attained using the proposed decoders is remarkably close to that of an optimal decoder, which uses impulse statistics; furthermore, the proposed decoders were compared with the alpha-penalty function decoder (alpha-PFD). The reported result showed that the proposed decoders are superior to an alpha-PFD because alpha-PFD neglects the channel memory property and experiences an error floor, in fairly general circumstances.