Return to search

Turbo coding and equalization for wireless communication systems

Turbo coding, a forward error correcting coding (FEC) technique, has made near Shannon Limit performance possible when Iterative decoding algorithms are used. Intersymbol interference (ISI) is a major problem in communication systems when information is transmitted through a wireless channel. Conventional approaches implement an equalizer to remove the ISI, but significant performance gain can be achieved through joint equalization and decoding. In this thesis, the suitability of turbo equalization as a means of achieving low bit error rate for high data communication systems over channels with intersymbol interference was investigated. A modified decision feedback equalizer algorithm (DFE) that provides significant improvement when compared with the conventional DFE is proposed. It estimates the data using the a priori information from the SISO channel decoder and also a priori detected data from previous iteration to minimize error propagation. Investigation was also carried out with Iterative decoding with imperfect minimum mean square error (MMSE) decision feedback equalizer, assuming soft outputs from the channel decoder that are independent identically distributed Gaussian random variables. The prefiltering method is considered in this thesis, where an all-pass filter is employed at the receiver before equalization to create a minimum phase overall impulse response. The band limited channel suffers performance degradation due to impulsive noise generated by electrical appliances. This thesis analysed a set of filter design criteria based on minimizing the bit error probability of impulse noise using digital smear filter.

Identiferoai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:616532
Date January 2013
CreatorsOletu, Grace Ogheneruonano
ContributorsRapajic, Predrag; Wu, Ruiheng
PublisherUniversity of Greenwich
Source SetsEthos UK
Detected LanguageEnglish
TypeElectronic Thesis or Dissertation
Sourcehttp://gala.gre.ac.uk/11626/

Page generated in 0.0015 seconds