DocumentCode
697977
Title
Decision diagram based computation of linearly independent ternary arithmetic transform spectra
Author
Lozano, Cicilia C. ; Falkowski, Bogdan J. ; Luba, Tadeusz
Author_Institution
Sch. of Electr. & Electron. Eng., Nanyang Technol. Univ., Singapore, Singapore
fYear
2009
fDate
24-28 Aug. 2009
Firstpage
973
Lastpage
977
Abstract
Classes of fastest linearly independent ternary arithmetic (FLITA) expansions have been proposed recently. They operate in arithmetic domain and have been shown to be useful for optimization of ternary functions representation. All FLITA transforms possess fast forward and inverse transforms and therefore can be calculated by fast transform method. However, it has been shown that for manipulation of large functions it is more advantageous to start from decision diagrams rather than truth vector. Hence in this paper new algorithm to obtain FLITA spectrum from ternary decision diagram is presented. It is developed based on the new notations for spectrum of an FLITA transform introduced here. The algorithm derives each spectral coefficient independently from each other, allowing the coefficients to be calculated in parallel manner. By starting from decision diagram, the algorithm enables the FLITA expansion to be computed for large functions for which the fast transform based algorithm may fail.
Keywords
decision diagrams; inverse transforms; optimisation; spectral analysis; FLITA expansions; arithmetic domain; decision diagram based computation; fast forward transforms; fastest linearly independent ternary arithmetic expansions; inverse transforms; linearly independent ternary arithmetic transform spectra; ternary decision diagram; ternary functions representation optimization; Boolean functions; Data structures; Laplace equations; Multivalued logic; Signal processing algorithms; Transforms; Vectors;
fLanguage
English
Publisher
ieee
Conference_Titel
Signal Processing Conference, 2009 17th European
Conference_Location
Glasgow
Print_ISBN
978-161-7388-76-7
Type
conf
Filename
7077550
Link To Document