9085

8430

كارشناسي ارشد

رياضي كاربردي

اصفهان: دانشگاه صنعتي اصفهان، دانشكده علوم رياضي

1392

يازده،97ص.: مصور،جدول،نمودار

ص.ع.به فارسي و انگليسي

مرتضي اسماعيلي

علي زاغيان

10/4/93

محمد حسام تدين، حميدرضا مرزبان

رياضي

ID8430

Deriving Good LDPC Convolutional Codes from LDPC Block Codes Mohsen Goodarzi m goodarzi@math iut ac ir 2014 Department of Mathematical Sciences Isfahan University of Technology Isfahan 84156 83111 Iran Supervisor Dr Morteza Esmaeili emorteza@cc iut ac ir Advisor Dr Ali Zaghian ali zaghian1338@yahoo com 2010 MSC 94BOS 68P30 Keywords convolutional codes low density parity check LDPC codes unwrapping graph cover AbstractLDPC convolutional codes are de ned by sparse parity check matrices analogous to LDPC blockcodes and they can be decoded by iterative message passing algorithms LDPC convolutional codeshave the capability of attaining good e ciency Two main methods have been introduced for makingthese codes The rst construction method has been introduced by Michael Tanner 33 This methodis based on parity check matrix of a quasi cyclic LDPC block code The second method has beenintroduced by Jimenez Feltstorm and Zigangirov 11 This method is based on unwrapping of paritycheck matrices of LDPC block codes In this thesis a code consruction method called graph covermethod is introduced and an algebraic representation of it is provided Using graph cover we study therelationship between two basic construction methods for LDPC convolutional codes Using the sameapproach a method based on graph cover is presented for deriving time varying LDPC convolutionalcodes from LDPC block codes Some LDPC convolutional codes resulted in from this method haveremarkable e ciency comparing to the underlying LDPC block codes This e ciency improvement LDPC convolutional codes compared to background LDPC block codes is called convolutionalgain Some aspects of this gain are reviewed here

مرتضي اسماعيلي

علي زاغيان

محمد حسام تدين، حميدرضا مرزبان